You are on page 1of 158

Tensional processes and perception of form

in three selected compositions by Kaija Saariaho

by Brad Jenkins

A 90-point exegesis and portfolio of original compositions

submitted to Massey University and Victoria University of Wellington
in partial fulfillment of the requirements for the degree
Master of Musical Arts (Composition)

New Zealand School of Music

! !
! !
! !

The primary feature that gives spectral music its stylistic uniqueness within the
field of art music is the blurring of the traditionally distinct roles of harmony and
timbre, through the use of chords derived from the naturally occurring overtones of
instrumental timbre (often referred to as timbre chords). Development of these chords
typically occurs very gradually, meaning it is often difficult to perceive the overall
form of a spectral work based on the progression through its constituent timbre
chords. This approach contrasts with the traditional reliance in both art music and
other Western music styles on perceivable pitch-based development as a primary
means of providing musical tension and form. Composers of spectral music must
rely on the manipulation and development of other musical parameters to provide
sufficient interest through foreground ornamentation while its underlying
harmonic/timbral macrostructure unfolds beneath.

This analysis shows how key musical parameters are manipulated over time to
provide tension and resolution (or, in Wallace Berrys terminology, progressive and
recessive processes 1 ), giving spectral works a perceivable, dynamic form.
Parameters examined include rate of harmonic change, dynamics, spectral/registral
spread, rhythmic activity, sound/noise, spectral density and
harmonicity/inharmonicity (the latter two providing a spectral analogue to
conventional notions of dissonance). Particular focus is placed on the rate of
harmonic change in the selected works and changes in the
harmonicity/inharmonicity (through spectral distortion) of harmonic material that
give spectral music its distinctive harmonic character. The way in which these
parameter curves intersect with one another is also examined.

For this study, three works by Finnish composer Kaija Saariaho are analysed. The
works cover a range of forces and display varying degrees of overtly spectral
influence: Nympha (1987) for string quartet and electronics, Du Cristal (1990) for
orchestra, and Cendres (1998) for piano, cello and flute. Analysis of the background
levels of parametric change reveal how Saariaho manages to maintain
microstructural interest in her spectral works while adhering to an underlying
macrostructural plan. Findings from this analysis will also be discussed in relation to
how they have influenced my own creative output for my MMA portfolio.

! 1!Berry, Wallace. Structural Functions in Music. Englewood Cliffs, N.J: Prentice-Hall,

1976. 11
! !

Firstly, I must thank my supervisor Michael Norris for his multifaceted guidance
over the course of my Masters studies, which has included proofing my drafts,
helping me to shape my stream of words into more coherent prose, technological
troubleshooting, providing solid leads on useful related research and giving valuable
feedback on my portfolio works. All of this helped me to leave our meetings feeling a
little more confident than when I arrived.

Thanks also to the musicians who gave readings of my portfolio works, the
recordings of which are enclosed with this exegesis: the New Zealand String Quartet
(Helene Pohl, Douglas Beilman, Gillian Ansell, Rolf Gjelsten), guitarist Jamie Garrick
and the NZSM Orchestra.

I would also like to acknowledge the New Zealand School of Music for offering me
an NZSM Scholarship for my Masters studies and thank the administrative staff at
NZSM for their support.

Last but certainly not least, big time gratitude to my wife Liz for her love and
patience while her sleep-deprived husband spent almost every waking minute of his
at-home time toiling away on a large quantity of words and music. Luv u babe xxx

This exegesis and portfolio is dedicated to Royce Galloway (Nana G) and Evelyn
Jenkins (Nana J), who both passed away during my studies.




1 Introduction

4 Methodology

12 Previous and Related Research

Analysis of works

17 Nympha

49 Du cristal

67 Cendres

83 Analysis summary

91 Relating research findings to my creative work

97 Bibliography

99 Discography


101 Original compositions for creative portfolio

148 CD tracklisting

CD containing recordings of portfolio compositions enclosed on inside back


At the heart of this exegesis is an exploration of how Saariaho succeeds in creating

tension and release in three selected works Nympha, Du cristal and Cendres. This
is of particular interest to me due to Saariahos avoidance of more conventional
tensional devices usually used in 20th century tonal music. Without goal-oriented
harmonic tension based on consonance and dissonance in equal-temperament,
Saariaho substitutes it with changes in harmonicity/inharmonicity based on an
underlying harmonic field, as well as with timbral transitions between noisy and
pure sounds. This analysis also looks at how form is defined in these works
through the use of either tensional trajectories or tensional stasis to produce an
audible sectionalisation of each work, revealing their formal structures. Selected
musical parameters are analysed to ascertain which of these parameters or
combinations of parameters dominate tensional processes in these works.

The decision to focus on the work of Saariaho is the result of my own interest, as a
composer, in spectral music. Throughout my undergraduate and postgraduate
studies in composition, I have been most intrigued by the music of spectral
composers such as Saariaho, Grard Grisey and Tristan Murail. I have been drawn to
the liminal harmonic states and timbral detailing that typify quintessential works of
this compositional approach. I am attracted to the organic movement of harmonic
material as it develops in spectral music, with pitch freed from the grid of the equal-
tempered chromatic scale. The use of quarter-tones and indefinite playing techniques
(such as glissandi) allow pitch and, by extension, harmony to occupy this
liminal space, allowing for the endless harmonic configurations that lie outside of
traditional harmonic structures of common-practice tonality. The process-based
approach to harmonic and timbral development can produce results that I may not
have consciously thought of otherwise.

I also find the focus on timbral manipulation in spectral music interesting for the
sense of internal development it conveys (by internal development I refer to the
manipulation of instrumental timbre so that certain partials vary in amplitude over
the duration of the sound, as opposed to discrete external development such as
changes in dynamics or pitch, or contrapuntal movement). I am interested in the
potential for the development of timbral detail to become part of short or long-term
teleological processes, or both. Historically, timbre (and the extended techniques

! 1!
associated with its manipulation) has been the musical parameter that has been
slowest in its incorporation into Western musical notation, presently integrated in as
elegant a manner as possible alongside the long-established noteheads, beams,
tempo markings, articulations, time signatures, key signatures and staff lines. Yet
timbre can exert a great deal of command over tensional states. Timbral tension is
capable of sharing a close link with physical labour and strain (as demonstrated, for
example, by the string playing in Nympha), just as high amplitude or a fast tempo
does. In musical contexts where parametric processes unfold slowly, timbral
transformation conveys a sense of microstructural life and activity. Spectral musics
general avoidance of conventional motivic material that slots into a metric and equal-
temperament grid allows the listener to be drawn to microstructural detail such as
timbral development. Motif is subordinated, existing in the service of texture but
occasionally drawn out as short-term ornamentation.

Though spectral composers acknowledge that the genre possesses a broad stylistic
identity (insisting that the process of generating pitch material through analysis of
sound spectra is generally the common feature, though there are still other possible
spectral approaches besides analysis-based processes), a connection exists between
the timbral detailing within spectral musics globalized, mass-oriented
orchestrational techniques and the contrapuntal saturation found in the works of
composers such as Varse, Ligeti, Penderecki and Xenakis in the decades preceding
the establishment of spectral music in the 1970s. This proto-spectralist music shares
with spectral music the organization of the ensemble into a sound mass, with an
individual line occasionally being lifted above the contrapuntal web to provide a
flicker of foreground interest. Looking further back we find the work of composers
such as Ravel and Debussy, who prioritized colour in their ensemble textures, an
early manifestation of the preoccupation with fused, resonant textures and stacked
tertian structures of many spectral composers. Over the past few decades, these
tertian structures have evolved to encompass quarter-tones that ordinarily would not
blend with equal-tempered pitches, but find fusion within the naturally occurring
harmonic spectra that found favour with early spectral composers.

Nympha, Du cristal and Cendres were selected for analysis due to their variety in
ensemble size, instrumentation and their exhibition, to varying degrees, of the
compositional techniques outlined in Saariahos own research2. The decade-long

2!Saariaho, Kaija. Timbre and Harmony: Interpolations of Timbral Structures.

Contemporary Music Review 2, no. 1 (1987): 93133. doi:10.1080/07494468708567055.!

! 2!
span (1987-1998) in which these works were written allows for a glimpse at how the
influence of spectral composition techniques changed over the course of time in
Saariahos writing and can also reveal any enduring parametric changes that
Saariaho calls upon to achieve certain tensional states. Indeed, by the late 1980s,
Saariaho and her contemporaries were considered post-spectralist. Metzer refers to
Du cristal as Saariahos last work in a period preoccupied with spectral and harmonic

The methodology developed for this analysis involves tracing the movement of
seven key parameters over the course of each work, as they move between states
associated with low or high tension. The parameters are:
rate of harmonic change
spectral/registral spread
rhythmic activity
spectral/harmonic density

Values are assigned to each parameter and scaled to reflect their relative dominance
or subordination alongside other parameters at points throughout each work.
Previous research related to this analysis is addressed, with particular reference to
Saariahos discussion of her parametricised approach to composition at the time of
writing. Relevant research regarding both musical tension and parametric interaction
is also addressed. Other existing analyses of the selected works are acknowledged,
with the differences in the focus of each analysis outlined.

The analyses of Nympha, Du cristal and Cendres are presented separately, with a final
analysis summary drawing together features shared by the three works regarding
the creation of tensional change, overall form and the use of foreground
ornamentation. Following the analysis section is a brief discussion of how Saariahos
approach to parametric interaction has influenced the composition of the works in
my own creative portfolio. More generalised aspects of spectral composition that
have been incorporated into my portfolio works are also described.
! 3!Metzer, David. Musical Modernism at the Turn of the Twenty-First Century.

Cambridge, New York: Cambridge University Press, 2009. 183


! 3!

Seven musical parameters have been measured: dynamic profile, registral spread,
harmonic density, rate of harmonic change, dominance of sound or noise, rhythmic
activity and degree of harmonicity/inharmonicity. Measurements are taken at both
the first beat of the bar and the downbeat closest to the middle of the bar, or only at
the first beat of the bar where there is little or no parametric change between these
two points.

The musical parameter that is most complicated in its measurement is harmonicity. I

chose to base my calculations on Hurons4 (1994) interval class index of 12-tone
equal-temperament consonance (Fig 1.1). Hurons index can be regarded as a rough
approximation of the perceived consonance of typical equal-tempered interval
classes. Huron does not provide a definition of consonance or dissonance but
remarks that the perception of an intervals consonance or dissonance depends on
the spectral content of the participating tones, their sound pressure levels and their
pitch register5. While sound pressure (dynamics) and registral spread feature as
measured parameters in this analysis, their effect on the perception of
harmonicity/inharmonicity in the selected works generally lies outside my analytical
scope. Instead, I will abstract away these complex details to preserve analytical
simplicity, relying purely on the consonance/dissonance values Huron ascribes to
each interval class as a measurement of total harmonicity/inharmonicity.

Hurons interval class index provides an index of numerical values that are applied
to each interval class in Fortes six-element interval vector template. Complementary
interval classes are given the same numerical value in Hurons index, as Fortes
interval vector inverts interval classes larger than a tritone. The values in Hurons
index are derived from results of three important studies in the perception of
consonance and dissonance: Malmberg (1918), Kameoka and Kuriyagawa (1969), and
Hutchinson and Knopoff (1979). The values established in each study were
normalized, with each set of values having a mean of zero and a standard deviation
of one. These previous studies apply individual index values to all 12 tones of the
octave, whereas Huron applied the same numerical value to each set of interval

! 4!Huron, David. Interval-Class Content in Equally Tempered Pitch-Class Sets:

Common Scales Exhibit Optimum Tonal Consonance. Music Perception: An

Interdisciplinary Journal 11, no. 3 (April 1, 1994): 289305. doi:10.2307/40285624.
5!ibid. 291!

! 4!
classes. An interval class with a consonance value higher than zero is considered
consonant, while an interval class with a consonance value lower than zero is
considered dissonant.

Interval class Consonance value

m2/M7 -1.428

M2/m7 -0.582

m3/M6 +0.594

M3/m6 +0.386

P4/P5 +1.24

A4/d5 -0.453

Fig. 1.1. Hurons Interval Class Index of Tonal Consonance6

Hurons index, however, covers only 12-tone equal temperament, making it

insufficient for measuring chords based on 24-tone equal temperament (found most
frequently in Nympha and Du cristal). To address this shortcoming, I interpolated
new index values for quarter-tone intervals by referring to the Plomp-Levelt curve7
(Fig. 1.2). One of the main reasons for employing the Plomp-Levelt curve in this
analysis is because of the continuous values provided between semitones (in contrast
to the discrete values in Hurons index), providing an approximate suggestion for
index values for quarter-tone intervals. As with other consonance and dissonance
indexes, the Plomp-Levelt curve is a model of the changing roughness, or
dissonance, of an interval between two pitches depending on their ratio-based
relationship. As the frequency interval increases, the interval is perceived as
increasingly dissonant, though only up to a certain distance known as the critical
bandwidth at which point perceived dissonance decreases. While Huron does not

! 6!Huron, David. Interval-Class Content in Equally Tempered Pitch-Class Sets:

Common Scales Exhibit Optimum Tonal Consonance. Music Perception: An

Interdisciplinary Journal 11, no. 3 (April 1, 1994): 289305. doi:10.2307/40285624.
7 Plomp, R., and Levelt, W.J.M. Tonal Consonance and Critical Bandwidth. Journal
of the Acoustical Society of America 38 (1965): 548560.!

! 5!
provide a definition of consonance or dissonance, Plomp and Levelt define the
phenomena of perceived consonance or dissonance of intervals as the result of
varying degrees of beating between adjacent frequencies, dependent on their
frequency ratio. Beating is the audible result of small differences in the cycle speed of
two or more pitch frequencies sounding simultaneously, and is the phenomenon to
which the description of roughness is applied regarding the perception of
dissonant intervals. Frequency ratios consisting of higher integers (e.g., 101:97) are
generally perceived as being more dissonant than those containing smaller, more
simple integers (e.g., 5:4). Plomp and Levelt produced a more detailed version of the
curve to accommodate the perception of intervallic consonance between complex
tones (tones containing harmonic overtone spectra such as those produced by
musical instruments, as opposed to sine waves, which do not have an overtone
series). This shows a general increase in the perception of consonance for intervals
that map closely to the most simple, low-integer frequency ratios. The index values
established in this analysis for quarter-tone intervals aim to reflect their general
tendency to sound rougher relative to their neighbouring equal-tempered
intervals, with some exceptions while the intervals of 12-tone equal temperament
map with sufficient closeness onto the simple just-intonation ratios to mostly be
considered consonant, the majority of quarter-tones are often located in the higher-
integer areas between the simpler just intonation points.

Fig. 1.2. The Plomp-Levelt curve for perception of intervallic consonance/dissonance, with
incorporation of data for perception of consonance/dissonance of interval classes between
two theoretical complex tones8

8!Plomp, R., and Levelt, W.J.M. Tonal Consonance and Critical Bandwidth. Journal

of the Acoustical Society of America 38 (1965): 548560., reproduced in Sethares, W. A.

Relating Tuning and Timbre. Experimental Musical Instruments 9, no. 2 (1993): 23

! 6!
In order to use the Plomp-Levelt curve to interpolate consonance values for 24-tet
intervals from their 12-tet neighbours, I first mapped the existing Huron interval
class index values onto the corresponding points for the 12-tet intervals on the curve.
For example, the m3/M6 value 0.594 from Hurons index was applied to that
intervals y-axis (sensory dissonance) position on the Plomp-Levelt graph, and so on
for Hurons other 12-tet interval values. This provided some approximate numerical
context for conferring consonance values to the curve points that corresponded with
quarter-tone intervals Plomp and Levelt provided either no or arbitrary numerical

Interval class Consonance value

(quarter-tone consonance values in
bold derived from P-V C)

m2 /M7 -1.7

m2/M7 -1.428

m2 /M7 -0.9

M2/m7 -0.582

M2 /m7 -0.5

m3/M6 +0.594

m3 /M6 -0.4

M3/m6 +0.386

M3 /m6 -0.4

P4/P5 +1.24

P4 /P5 -0.4

A4/d5 -0.453

Figure. 1.3 Interval class index used by author for measuring harmonicity/inharmonicity.
Quarter-tones derived through Plomp-Levelt curve interpolation. Semitones derived from
Hurons interval class index.

! 7!
values for the y-axis of their graph. The lowest y-axis value is 1.24 (P4/P5), at almost
the lowest possible point of the y-axis, with the highest being greater than -1.428
(since the y-axis is measuring sensory dissonance, Hurons index values are
obviously put in reverse numerical order on the Plomp-Levelt y-axis). The resulting
interval class index, which covers a 24-tet octave, is shown in Figure 1.3.

The following is a demonstration of the harmonicity calculation process for harmonic

material in the three analysed works:

Fig. 1.4. Sample chord for calculation of harmonicity/inharmonicity, with constituent 12
and 24-tet intervals (chord example devised by author)

The consonance value of each interval within the sample chord (Fig. 1.4) is noted,
with these values then added together to give a total consonance value (Fig. 1.5):

Pitch Interval Consonance

classes name value

a. E A# A4/d5 -0.453

b. EC M3/m6 +0.594

c. EF m2 -1.7

d. A# C M2 -0.582

e. A# F P5 -0.4

f. CF P4 -0.4

Total = -2.941

Fig. 1.5. Process for obtaining consonance value of sample chord in Fig. 1.4.

The negative value obtained by combining the consonance values of each intervallic
relationship within the chord indicates that this combination of pitches is likely to be
perceived as more dissonant (containing more harmonic tension) than consonant.

! 8!
Harmonicity is the only parameter in this analysis that includes numerical values
less than zero. Indeed, much of the harmonic material in the selected works is
inharmonic according the standard of measurement applied here, meaning the
graphed data for this parameter regularly appears beneath the zero value on the y-
axis. To allow a clearer view of the relationship between harmonicity/inharmonicity
and other parameters, it is presented in differing graphical ways. In the sectional
analyses of Nympha and Cendres harmonicity/inharmonicity is presented on
separate graphs whose y-axes display both positive and negative values. Du cristal,
on the other hand, contains thick harmonic fields whose levels of
harmonicity/inharmonicity are rarely consonant. This allows
harmonicity/inharmonicity to appear on the same section graphs as other
parameters, with consonance value placed on a secondary y-axis with its values

Dynamic profile values have been obtained by analysing amplitude levels of the
works as .wav files and applying a value to the amplitude of each bar based on a
scale of zero to 10. A zero value reflects silence and a value of ten reflects maximum

Harmonic density values are obtained by measuring the number of different pitches
sounding at once. For example, if all players in a string quartet are performing
double-stopped dyads with no pitches doubled in unison among them, the
numerical value is 8. If two players are performing identical dyads the harmonic
density would then have a value of 6, as there are now only six pitches sounding at
once (Fig. 1.6). For each work the highest and lowest values of harmonic density are
scaled so that the maximum value for harmonic density is equal to 10. In this study, a
chord is defined as two or more different pitches sounding simultaneously.

Fig. 1.6. Process for measurement of harmonic density

(chord example devised by author)

Values for registral spread are obtained by counting the number of semitones
between the highest and lowest pitches in a chord. For example, a chord with a

! 9!
registral spread spanning three octaves would have a numerical value of 36. As with
harmonic density values for each work, the highest and lowest values of registral
spread are scaled so that the maximum value for harmonic density is equal to 10.

In assigning a numerical value to rhythmic activity, identifying the factors to be

taken into account for this parameter made for a slightly more intuitive process. The
primary criteria was the degree to which rhythmic activity acted as a foreground
feature, either through the line of one player or through two, three or four-part
contrapuntal material essentially the exposure of a perceivable underlying pulse,
and the importance of this device when it appears, relative to other parameters. A
zero value is applied where there is no rhythmic material present (or such a high
density of individual rhythmic material that no actual rhythmic activity is
perceivable) and a value of 10 is given where the work is dominated by highly
rhythmicized foreground material.

The dominant timbral quality of a bar is measured by referring to Saariahos own

sound-noise axis, which the composer has used as an alternative tensional device to
harmonic consonance and dissonance sound replacing harmonic consonance and
noise replacing dissonance 9 . Noise in which any sense of definable pitch is
obscured due to frequency saturation is given a value of 10, while purer sounds
(for example, a violin played legato, with naturale bow position and without
excessive bow pressure) are measured at zero.

To accurately calculate the rate of harmonic change (a change in the underlying

harmonic field from which combinations of pitches are extracted) in each work, clock
time is used as a measure of time as opposed to bar numbers, as musical time is of
course subject to tempo fluctuations. Each work is analysed in overlapping 30-
second segments (0:000:30, 0:150:45, 0:301:00 and so on), with the number of
changes in the underlying harmonic field recorded for each 30-second segment.

Aside from harmonicity/inharmonicity in Nympha and Cendres and rate of

harmonic change for all three works, all parameters are scaled between zero and 10,
allowing for multiple parameters to appear on the same section graph. The result of
this multi-parameter graphic representation is that section-based changes to
parametric dominance and relationships are more clearly displayed. With the

! 9 !Saariaho, Kaija. Timbre and Harmony: Interpolations of Timbral Structures.

Contemporary Music Review 2, no. 1 (1987): 93133. doi:10.1080/07494468708567055.

! 10!
exception of clock time for the rate of harmonic change and overall tensional
trajectory graphs, the x-axis value always shows bar numbers, and the y-axis the
scaled value for the tensional state of each parameter. As previously mentioned,
harmonicity/inharmonicity in Du Cristal appears on the same section graphs as other
parameters, with consonance value placed on a secondary y-axis with its values

In addition to the graphic representation of the seven parameters outlined here, a

general tensional trajectory line is superimposed over some graphs to clarify a
sections formal function. The values of the tensional trajectory lines are not based on
any calculation involving the dominant parameters in a given section. Rather, their
values are based on my experiential perception of the tensional state in each section
and is intended to be relative to the tensional trajectory across the entire work (the
tensional trajectory across the entirety of each work is also shown at the beginning of
each analysis).

The effect of chord spacing on the perception of harmonicity is not directly

addressed in this analysis. Danner (1985) explores the effect of spacing on the
perceived dissonance of a chord, finding that chords containing larger intervals
between their constituent pitches are generally perceived as being markedly less
dissonant than the same chord with its pitches more closely spaced 10 . This
phenomenon may affect the perception of chordal harmonicity at certain points in
the Saariaho works analysed here (especially given the prevalent inharmonicity of
the harmonic material), but in this case it is generally out of the scope of the analysis.
It can be confidently assumed that wide chord spacing has a leveling effect on the
perception of harmonicity. For example, if a relatively consonant chord such as a
major triad had its three pitches spread across four octaves, this would result in the
triad being perceived as not necessarily dissonant, but as being somewhat less
consonant than in its closer spacing. A clearer description of this effect could be that
wider chord spacing results in a generally more ambiguous sense of harmonicity.
! !

! 10!Danner, Gregory. The Use of Acoustic Measures of Dissonance to Characterize

Pitch-Class Sets. Music Perception: An Interdisciplinary Journal 3, no. 1 (October 1,

1985): 10322. doi:10.2307/40285324.

! 11!
Granot and Eitan (2011) have explored ways in which the interaction of selected
musical parameters contributes to the perception of musical tension. One of the
hypotheses of their research states that loudness is the musical parameter that has
the most powerful effect on the perception of musical tension. This hypothesis was
proved correct in their findings. In explaining their reasoning, the authors refer to
the effect of loudness in extramusical environments, in which perception of this
phenomenon serves as an instinctive biological warning. Changing loudness can
indicate whether a particular sound source is approaching or moving away, allowing
an organism to evaluate whether or not it is in danger. Granot and Eitan also
maintain that another reason for loudness being linked with tension in a musical
context is because it is the main auditory manifestation of physical force (2011:
241) 11 . This, of course, is also a sensation brought across from extramusical

Cohen and Wagner (2000) address parametric interaction in musical tensional

processes. One of the suggested reasons for the perception of increasing musical
tension is the presence of an inverse relationship between the extent to which
parametric changes correlate and the amount of information contained in the overall
change. The authors suggest that when all parameters synchronise that is, when
all parameters simultaneously move in the same direction between high and low
tension states the overall informational change is actually minimal, lessening the
overall amount of tension, and vice versa. The greater the amount of information, the
higher the cognitive load, which itself affects tension (Cohen and Wagner, 2000)12.
This analysis of Du cristal, Nympha and Cendres also looks to identify any instances
of this type of nonconcurrent parametric interaction within these works.

Dibben (1999) looks at the perception of structural stability in atonal music. The
results of two experiments provide insights that relate to questions concerning the

! 11!Granot, Roni Y., and Zohar Eitan. Musical Tension and the Interaction of

Dynamic Auditory Parameters. Music Perception: An Interdisciplinary Journal 28, no. 3

(February 1, 2011): 21946. doi:10.1525/mp.2011.28.3.219.
! 12!Cohen, Dalia, and Naphtali Wagner. Concurrence and Nonconcurrence between
Learned and Natural Schemata: The Case of J. S. Bachs Saraband in C Minor for
Cello Solo. Journal of New Music Research 29, no. 1 (March 2000): 23.

! 12!
perception of form in Du cristal, Nympha and Cendres. Based on experimental results,
Dibben suggests that listeners might hear atonal music in terms of the relative
structural importance of events and that listeners hearing is greatly influenced by
metrical and durational structure (Dibben, 1999)13. Dibben goes on to posit the idea
that salience, voice-leading and dissonance could be crucial elements that suggest
relationships inferring prolongational structure in atonal music. What Dibben refers
to as salience relates to the idea of foreground ornamentation mentioned earlier, a
foregrounded detail in one or more musical parameters that serves to maintain short-
term interest while underlying macrostructural processes continue to unfold

The main catalyst for this analysis comes from Saariahos own research and writing
on tensional processes and the perception of form. In writing about the development
of musical form, Saariaho (1987) has referred to Russian painter and art theorist
Vassily Kandinskys definition of form as the external manifestation of inner
meaning. Accordingly, this sense of inner meaning is realized by Saariaho at the
level of musical parameters instead of the level of a works overall formal structure.
One of the parameters addressed by Saariaho is her sound/noise axis (in which
noise replaces the concept of dissonance and sound that of consonance (1987:
93)) 14 , the composers focus on this necessitating its inclusion as a measured
parameter in this analysis. Saariaho explores the substitution of consonance and
dissonance with pure sounds and noise, respectively. In Du cristal, Nympha and
Cendres, Saariaho still relies on intervallic relationships of pitched material to
generate tension, though not in a tonal setting. However, the sound/noise axis is
overlaid as a means of augmenting the tensional possibilities of the works,
particularly in Nympha. Saariaho regards the form of a work as something that, in
the absence of overt harmonic rhythm, the listener perceives based on recognizable
landmarks. The listeners memory of the work is simplified, so that it is remembered
for its moments containing more salient features, as Dibben would describe them.
Saariaho cites her work Verblendungen (1982-84) as the first work in which she

! 13!Dibben, Nicola. The Perception of Structural Stability in Atonal Music: The

Influence of Salience, Stability, Horizontal Motion, Pitch Commonality, and

Dissonance. Music Perception: An Interdisciplinary Journal 16, no. 3 (April 1, 1999):
26594. doi:10.2307/40285794.
! 14 !Saariaho, Kaija. Timbre and Harmony: Interpolations of Timbral Structures.
Contemporary Music Review 2, no. 1 (1987): 93133. doi:10.1080/07494468708567055.

! 13!
attempted to construct forms of tension as well as structures of several levels15.
Trajectories were created for each parameter polyphony, rate of harmonic
progression, pitch range, dynamic profile, homophony, foregrounding of its tape
part with the intended result being a work whose form is derived from the
interaction of these parameters, with culminating points that serve as memorable
structural landmarks.

Howell (2011) has discussed Nympha as an example of how Saariaho controls

teleological listening and the processes by which the composer builds musical
narratives16. Regarding general formal structure, Howell maintains that Nympha can
be divided into two identifiable parts (separated by a double barline that is directly
preceded by a rare moment of silence). Howell then offers a more detailed
breakdown of the formal structure of each part, describing them as the products of a
series of developmental cycles, loosely fractal in nature, which the author compares
to the action of a water lily opening (Nympha means water lily). In keeping with the
botanical theme of the work, Hargreaves labels the developmental cycles as growth-
death-decay cycles and maintains that, as the work progresses, the cycles vary in
duration, harmonicity, and their relationship to adjacent cycles, with some cycles
joined together or appearing within another, larger cycle. The basis of my analysis
differs from Howells through its more linear division of the work into discrete
sections (no overlapping sections) defined by tensional trajectories and states, as well
as parametric combinations, which illuminate the form of the work.

Several analyses of Du cristal exist, with the perception of form being a common
topic. The degree of depth to which the work is explored varies from author to
author, with an unsurprising variety of conclusions regarding its formal structure.
This variety has much to do with the difference in criteria by which form is judged in
each analysis.

Hargreaves (2011) analysis of Du cristal focuses on the musical, semiotic and

philosophical relationships that exist within the work. Echoing Saariahos thoughts
on formal structure being perceived through the identification of culminating
points of musical parameters, Hargreaves describes the formal structure of Du

! 15 !Saariaho, Kaija. Timbre and Harmony: Interpolations of Timbral Structures.

Contemporary Music Review 2, no. 1 (1987): 93133. doi:10.1080/07494468708567055.

! 16!Howell, Tim, Jon Hargreaves, and Michael Rofe. Kaija Saariaho: Visions, Narratives,
Dialogues. Burlington, VT, Farnham, Surrey: Ashgate, 2011. 81105

! 14!
cristal as an uninterrupted 16-minute sound, different parts of which come to
prominence at different times (2011:184)17. Hargreaves goes on to outline two ways
in which Saariaho complicates the process of perceiving form in Du cristal. Firstly,
Saariahos simultaneous manipulation of multiple musical parameters (describing
them as being in an ongoing competition for prevalence18) means that the form of
the work is not realized through one clear parametric narrative. Secondly, the
development of textural features in Du cristals background material is such that it
spreads thoroughly throughout much of this layer, obscuring larger structural
patterns that lie beneath. Hargreaves cites Saariahos scheme for pitch organization
in Du cristal, namely regarding pitch distribution and density, as an effective device
for controlling the sense of tension and release during the work. The parameters
Hargreaves focuses on are tempo, pitch and textural patterns, while also addressing
Saariahos casting of the orchestras sections into register-defined bands. Hargreaves
does not go so far as positing his own formal breakdown of Du cristal, instead
outlining analyses by Metzer and Pousset alongside a description of events that
take place over the works duration.

Pousset stresses the repetitive qualities of Du cristal as a key to revealing its form,
though we are not dealing, of course, with traditional literal repetition. Alluding to
the symmetrical nature of the works namesake, Pousset considers Du cristal to be
organised in a ranging ternary form A B A preceded by a slow introduction,
with the second A section consisting of mutations of material found in the original A
section, now ordered in retrograde19. While acknowledging the traditional role of
tension and release in the continuation of forward movement in Du cristal (this being
the main point of difference between Poussets analysis and my own the focus on
tension and release as Du cristals primary formal device), Pousset cites repetition
and equilibrium as the most important criteria by which to consider the formal
coherence of Du cristal. Poussets additional acknowledgment, like Hargreaves, of
what Stoanova (1994) describes as the network of internal relations and temporal

! 17!Howell, Tim, Jon Hargreaves, and Michael Rofe. Kaija Saariaho: Visions, Narratives,

Dialogues. Burlington, VT, Farnham, Surrey: Ashgate, 2011. 177201

18!ibid. 192!
! 19!Pousset, Damien, Joshua Fineberg, and Ronan Hyacinthe. The Works of Kaija
Saariaho, Philippe Hurel and Marc-Andr DalbavieStile Concertato, Stile
Concitato, Stile Rappresentativo. Contemporary Music Review 19, no. 3 (January 1,
2000): 100101

! 15!
equilibria of sonic materials and processes20 in Du cristal perhaps accounts for the
relative simplicity of his formal breakdown of the work, as a more detailed foray into
its formal intricacies based on repetition and equilibrium may prove too

Metzer restricts his focus to the opening and closing sections of Du cristal, but also
regards the work as consisting of three sections (A = bb.1126, B = bb.127299, C =
bb.300373). Metzer judges the form of Du cristal through two focal points: the
raising of the solo voices and how the voices assume expressive presence through the
medium of flux21. Stating that Du cristal contains recurring passages and sounds,
Metzer suggests that the form of the work is articulated by the variations based on
this recurring material. While Saariahos pre-compositional sketches may indicate
the repetition of material, it seems that very little of this repeated material is
organized in such way as to be consciously perceived. This is where the value of a
tension-based approach to defining form is useful, as the experiential perception of
tension seems a more tangible route to understanding its structure. Focusing on
parametric interactions and dominance further clarifies form, revealing how they
combine over shorter durations to create tensional trajectories or states. Perception is,
of course, unique to each listener, so the subjectivity of experiential perception must
also be acknowledged.

20!Stoanova, Ivanka. A Work of Synthesis: Analysis of Amers 1994. cited in

Pousset. 101!
! 21!Metzer, David. Musical Modernism at the Turn of the Twenty-First Century.
Cambridge, New York: Cambridge University Press, 2009. 186

! 16!

for string quartet and electronics

This analysis discusses tensional processes employed in Saariahos Nympha for

string quartet and electronics by tracing changes to each of the parameters
introduced in the Methodology section. Tension and release are used as guides for
dividing the work into distinct sections, where the combined parametric changes
suggest localised beginnings and endings that contribute to a sense of overall form.
Important relationships between specific parameters are highlighted, as are
Saariahos methods of subverting expectations in the creation of tension and release
by supressing or accentuating certain parameters. Saariahos use of varying degrees
of parametric desynchronisation particularly with those parameters frequently
used to provide foreground ornamentation in Nympha are also discussed.

As expected, the rate of harmonic change in Nympha is mostly rather slow, though
there are more active fluctuations during several passages throughout the work (Fig.
2.2). In particular, the rate of harmonic change can be seen to increase during the two
most climactic passages in Nympha (bb.151207 and bb.296334), with these
increases roughly lining up with increases in dynamic profile. An increase in the rate
of harmonic change is a common tensional device employed in many forms of
Western music, and it is interesting to find this device employed by Saariaho in this
art music setting, especially since other foregrounded parameters such as
sound/noise can obscure the rate of harmonic change in Nympha. For example,
Saariahos use of playing techniques such as extreme bow pressure means that any
discernible pitches are buried beneath the resulting noise, with extended passages of
extreme bow pressure sometimes masking the underlying accelerated harmonic

The sound/noise parameter is key in controlling tension and release in Nympha. As

Nympha navigates between pure tones and noise, it becomes apparent that it is
almost never without some form of timbral manipulation adding a degree of noise to
its overall sound. This can be seen in the restless bowing instructions prescribed for
each player in the quartet, which constantly oscillates between sul tasto and sul
ponticello bow positions. As a result, the timbral foundation of Nympha is infused
with, at the very least, a small but near-constant amount of noise.

Rhythmic activity and sound/noise are the two parameters that are most

! 17!
consistently foregrounded, providing the moment-to-moment foreground
ornamentation, while also responsible for expressing larger synchronous and
tensional changes, in Nympha. While harmonic development unfolds beneath, the
interaction of these two parameters, especially in the first half of the work, seems to
be one of alternating dominance of the foreground.

For instance, following the unison A that opens Nympha, a movement to the noisy
end of the sound/noise continuum occurs, as all players employ surges of extreme
bow pressure. From b.36, a clear polyrhythmic passage emerges, which then
refocuses into a unified group of semiquavers at b.39. A hocketed rhythmic passage
follows, centred microtonally around a unison G4, with the duration of some notes
slightly extended. By the time the rhythmic material assumes the foreground, the
noise content of the material has significantly receded due to the absence of extreme
bow pressure, leaving the variations in bow position as the only noise-creating
techniques. By b.50, rhythmic clarity has dissolved and noise has returned as the
foregrounded feature, with extreme bow pressure used in the first violin part once

The second climax of Nympha takes place when the foregrounding of the rhythmic
and sound/noise parameters are combined, which is significant for its rarity in
Nympha and thus structurally important. At b.326, periodic slabs of extreme bow
pressure are regularly passed around the quartet. Within the sustained wall of noise,
each down-bow is audible, organized into a hocketed pulse. In terms of providing
formal definition, this amalgamation of the two parameters serves as the eventual
realization of something that had been briefly touched upon, but largely evaded up
to this point the combining of tensional high-points of rhythmic activity and
sound/noise. The only other comparable amalgamation of these parameters occurs
from bb.195207, in which the quartet unites in rhythmic, accented stabs of extreme
bow pressure. Harmonic parameters harmonicity, registral spread, rate of
harmonic change are rendered temporarily irrelevant by the dominance of the
sound/noise parameter being pushed to its extreme, especially in the latter of these
two events, where the bow pressure is such that the notated pitches are barely
audible. Much of the structural importance of the passages overwhelmingly
dominated by noise comes from the tensional extreme they inhabit. As discussed in
the Methodology, noises saturation of frequencies goes beyond mere chromatic
saturation, with much of the tensional impact derived from an implied loss of control
over the production of sound, especially in combination with high amplitude, which
is also a parametric feature of Nymphas climactic material.

! 18!
The harmonicity and harmonic density parameters are closely related for much of
Nympha, with inharmonicity often increasing as harmonic density increases. This
relationship is a fairly predictable one, given that the number of different intervallic
relationships within a chord is likely to increase as the number of pitch classes in that
chord increases. Importantly, only the minor third, major third and perfect fourth
intervals are given positive numerical, or consonance, values in Hurons interval
class index. Not surprisingly, the vast majority of Nymphas harmonic content is
inharmonic. A strong contributor to the degree of inharmonicity in Nympha is the
presence of quarter-tones, which have a high dissonance value in relation to equal-
tempered intervals. In scaling the harmonicity values for graphical representation
alongside other musical parameters, this parameter is graphed using the secondary
y-axis, with the values on this axis reversed so that high-points in its graphical
representation indicate an increased degree of harmonic tension (or values that are
below zero according to Hurons index), since this analysis is concerned with the
modelling and comparison of tension through multiple parameters. Here, the scaled
zero value on the secondary axis is 7.84.

Harmonic density and inharmonicity rises and falls as Saariaho extracts varying
combinations of pitches from the underlying harmonic fields. Harmonic tension rises
and recedes in waves.

Aside from being perceivable in recorded performances, a sense of how Saariaho

constructs form in Nympha can also be gained from following changes to the musical
parameters charted in this analysis. As Howell (2011) has noted, Nympha is, at the
most basic level, divided into two sections. This is indicated in the score by the
double barline at the end of b.262, the only double barline to appear within the score.

SECTION 1 : bb.124
The first section of Nympha introduces the sonic qualities that generally underpin
the work: sustained pitches, with each part containing independent surges in
dynamic profile and constantly changing bow position and pressure; a harmonic
character that includes quarter-tones, with moments of rhythmic, proto-melodic
material dominating the works surface. For the purposes of this analysis, the term
proto-melodic is used to describe a foregrounded sequence of pitches that lacks most
of the properties one would associate with a melodic line. In particular, a proto-
melodic line 1) possesses little or no variety in pitch contour, 2) employs little
rhythmic activity (that is, generates little rhythmic tension around an underlying

! 19!
Section Bars Time Tensional Main features

1 124 0:001:16 low tension Introduces independent surges in

dynamic profile and constantly
changing bow position/pressure;
gradual shift in dominant foreground
ornamentation parameter
(sound/noise to rhythmic activity)

2 2538 1:171:50 mid tension Louder dynamics and change in

attack profile, appearance first
muted climax chord with wide
registral spread, then narrowing to
single pitch

3 3962 1:512:26 mid to high tension first instance of rhythmic unison,

return of extreme bow pressure

4 63104 2:273:59 low tension rest area, more parametric stability

5 105150 4:005:28 mid tension to high First main climax wide registral
tension spread, dynamic unison, increased
rate of harmonic change, first use of
repetitive material, foregrounded
rhythmic activity

6 151207 5:297:35 high tension Continuation of climax more

unified playing, synchronized use of
extreme bow pressure, rhythmic
activity and sound/noise parameters
sharing the foreground (point of
arrival), increased rate of harmonic
change, but with contracted smaller
registral spread

7 208243 7:369:17 low tension Introduction of structurally

significant whispered syllabic
fragments, return to scattered bow

8 244262 9:1810:18 low tension Second appearance of muted

climax chord to conclude the first
half of Nympha, increased rate of
harmonic change but little tension

9 263295 10:1913:05 low to high tension Parametric surges within controlled

increase in tension, gradual
synchronization of dynamic surges,
absence of foreground ornamentation

! 20!
10 296334 13:0615:08 high to low tension Major climax rhythmic activity and
sound/noise parameters sharing the
foreground (point of arrival), high
amount of pitch instability

11 335362 15:0916:39 low tension Proto-melodic material as foreground

ornamentation, third appearance of
muted climax chord at end of

12 363389 16:4018:40 low tension Return of whispered material, shift in

dominant foreground ornamentation
parameter (rhythmic activity to

Fig. 2.1. Formal outline of Nympha


0.0" 1:00" 2:00" 3:00" 4:00" 5:00" 6:00" 7:00" 8:00" 9:00" 10:00" 11:00" 12:00" 13:00" 14:00" 15:00" 16:00" 17:00" 18:00"



0.0" 1:00" 2:00" 3:00" 4:00" 5:00" 6:00" 7:00" 8:00" 9:00" 10:00" 11:00" 12:00" 13:00" 14:00" 15:00" 16:00" 17:00" 18:00"


Fig. 2.2. Rate of harmonic change and overall tensional trajectory in Nympha. The overall
tensional trajectory shown here is simply Nymphas constituent sectional tensional
trajectories mapped end-to-end. Each sectional tensional trajectory is presented and
discussed in the analysis that follows, with reference to the musical parameters that
contribute most to that trajectory. All sectional tensional trajectories are based on the
authors subjective perception of tensional change, using data from the analysis of
parametric changes as a source of evidence. Graphs for Du cristal and Cendres have also
been produced by the same method, and appear later in this exegesis.

! 21!
pulse) and 3) produces little in the way of tonal tension and release with its sequence
of pitches.

The primary development in this section is a gradual shift in foreground material,

from arhythmic surges of extreme bow pressure on all instruments (Ex. 1.1) to a
proto-melodic line played in natural bow position by the second violin. The surges of
extreme bow pressure initially overlap so as to create an almost unbroken line of
noise amongst the purer tones. Across the first minute of the piece, however, the
surges of extreme bow pressure become less frequent. The decreasing presence of
extreme bow pressure occurs in contrast to the second violins proto-melodic
material (Fig. 2.3), which first emerges in b.8 as a three-note fragment, growing in
length with each iteration until it becomes an unaccompanied line by b.23.

Ex. 1.1 Use of extreme bow pressure (notated by black lines) in the first section of

! 22!

1" 2" 3" 4" 5" 6" 7" 8" 9" 10" 11" 12" 13" 14" 15" 16" 17" 18" 19" 20" 21" 22" 23" 24"
Sound/noise" Dynamic"pro<ile" Rhythmic"activity" Tensional"trajectory"

1$ 2$ 3$ 4$ 5$ 6$ 7$ 8$ 9$ 10$ 11$ 12$ 13$ 14$ 15$ 16$ 17$ 18$ 19$ 20$ 21$ 22$ 23$ 24$

Fig. 2.3. Parametric changes in the first section of Nympha.

SECTION 2 : bb.2538
The brief second section differentiates itself from the first in its use of parameters
other than noise for generating tension, and for containing one of Nymphas
strongest formal landmarks. Firstly, the attack profiles of the pitches now start at
peak amplitude and recede, unlike the first section, in which pitches mostly grow out
of, and recede back to, niente dynamics. The louder dynamics incorporated into the
change in attack profile demonstrate the first parametric transformation through
which Saariaho builds tension in this section (Fig. 2.4). The wash of pitches is once
again arhythmic, though extreme bow pressure is momentarily absent. The decrease
in tension from the stabilisation of the sound/noise parameter is somewhat
neutralised by the change in attack profile and dynamics, which creates a more
aggressive arhythmic character at the start of this section. Harmonic tension is also
greater than that of the first section.

! 23!

25" 26" 27" 28" 29" 30" 31" 32" 33" 34" 35" 36" 37" 38"
Dynamic"pro6ile" Harmonic"density" Registral"spread" Tensional"trajectory"

Fig. 2.4. Parametric changes in the second section of Nympha.

Ex. 1.2. First statement of C# chord, in the section section of Nympha.

The registral spread of the work expands from the confines of the middle-to-upper
registers of the opening section, with the cello descending to a low C#,! while the first
violin ascends to a C7! at b.31. This registral change coincides with changes in
harmonic density and dynamic profile. All instruments revert to playing a single
pitch (or a trill note in the case of the second violin) and the most noticeable change
in Nympha up to this point occurs: the dynamic profile drops to p, with the viola and
cello holding bowed tremolos whose sustained instability introduces a new

! 24!
tensional quality to the work. Pitches from a new harmonic field begin to emerge as
the quartet reaches the sustained, widely spread chord at b.31, with this being the
first of three moments in Nympha where this chord features (Ex. 1.2). While never

identical in its constituent pitches (here they are C#-B-G-C), it is distinguishable by

the cellos low C# played with a bowed tremolo. The purpose of this chord appears
to be cadential and is discussed later in this analysis as it makes further appearances.
This chords formal purpose as a cadential event is reinforced by the change in the
underlying harmonic field that also begins at this point. The registral spread
contracts to G4 and its surrounding quarter-tones in preparation for the growth of a
new harmonic field and new section of the work.!
SECTION 3 : bb.3962
The tensional trajectory of the third section is a moderate shift towards a higher state
of tension, culminating in a crescendo. The section begins with some of the strongest
foregrounding of rhythmic activity in the entire work, signalling another clear
structural landmark (Fig. 2.5). In this section, however, tension is closely tied to
registral spread, which is initially minimal, being centred around G4. Coupled with
moderate dynamics, the result is a moment of rhythmic unison whose tensional and
climactic potential is only partially fulfilled in its first appearance. This allows for
tensional intensification of rhythmic unison as it reappears in the work.
As the quartet falls into rhythmic unison at b.39 (Ex. 1.3), playing a G4 in even
semiquavers, there is an accompanying increase in tempo by way of metric
modulation. The foregrounding of rhythmic activity continues as the semiquaver
rhythm quickly splinters into a hocketed, syncopated rhythm. Quarter-tones
surrounding G4 also return as a precursor to the registral spread extending upward
and a more complex, arhythmic counterpoint returning that is reminiscent of the
opening material of the second section. However, this passage is more restless due to
its tempo fluctuations and the return of extreme bow pressure, which features in the
first violin part. Independent dynamic surges in each part gradually build to a
crescendo at b.62, at which point the tempo briefly drops back to lento, ushering in
the next section.

! 25!

39" 40" 41" 42" 43" 44" 45" 46" 47" 48" 49" 50" 51" 52" 53" 54" 55" 56" 57" 58" 59" 60" 61" 62"
Dynamic"pro6ile" Harmonic"density" Registral"spread" Tensional"trajectory"

39" 40" 41" 42" 43" 44" 45" 46" 47" 48" 49" 50" 51" 52" 53" 54" 55" 56" 57" 58" 59" 60" 61" 62"
Rhythmic"activity" Sound/noise"

39$ 40$ 41$ 42$ 43$ 44$ 45$ 46$ 47$ 48$ 49$ 50$ 51$ 52$ 53$ 54$ 55$ 56$ 57$ 58$ 59$ 60$ 61$ 62$

Fig. 2.5. Parametric changes in the third section of Nympha.

! 26!
Ex. 1.3. Rhythmic unison in the third section of Nympha.

SECTION 4 : bb.63104
Coming down from the crescendo, harmonic density increases from that of the
previous section, though the fourth section is notable for passages in which the
usually closely linked harmonic density and inharmonicity parameters diverge
somewhat. For example, between bb.7177 and bb.89105 the degree of
inharmonicity is relatively low despite a general increase in harmonic density from
the third section (Fig. 2.6). Here, inharmonicity veers a little either side of the zero
value of 7.84. Indeed, this section seems to develop into a type of rest area within
the work, containing relatively less tension.

The overall dynamic profile remains quite constant through this section, as does
registral spread. The arhythmic wash of pitches described in the second section
returns for much of the fourth section, with extreme bow pressure being
reintroduced as a prominent surface feature. Bow pressure is passed between
instruments, again scattered irregularly between the players. The perception of

! 27!

63" 65" 67" 69" 71" 73" 75" 77" 79" 81" 83" 85" 87" 89" 91" 93" 95" 97" 99" 101" 103"
Sound/noise" Dynamic"pro:ile" Harmonic"density" Tensional"trajectory"


63$ 65$ 67$ 69$ 71$ 73$ 75$ 77$ 79$ 81$ 83$ 85$ 87$ 89$ 91$ 93$ 95$ 97$ 99$ 101$ 103$


Fig. 2.6. Parametric changes in the fourth section of Nympha.

extreme bow pressure as a timbral motif in Nympha is reinforced through its

frequent use in passages such as this. The amount of pressure and the density with
which the technique is clustered between instruments in the work are major factors
in altering levels of tension, hence the aforementioned irregular scattering of extreme
bow pressure in this rest area. It is applied in a more forceful and clustered fashion
for highly tensional passages later in the work.

SECTION 5 : bb.105150
The fifth section is a climactic passage, with the climactic material realized through
both conventional devices found in classical music and through devices that are
significant in the context of Nymphas soundworld. The most obvious parametric
feature is the wide, cadential registral spread that remains throughout the section,
indicative of a climactic moment, with particular use of pitches in the bass register
suggesting a point of arrival (Fig. 2.7).

Following the crescendo at the end of the fourth section, all instruments are united at
a forte dynamic, the loudest point of the work thus far (all instruments playing at the
same dynamic simultaneously is rare up to this point in Nympha). This forte passage

! 28!

105" 107" 109" 111" 113" 115" 117" 119" 121" 123" 125" 127" 129" 131" 133" 135" 137" 139" 141" 143" 145" 147" 149"
Dynamic"pro4ile" Registral"spread" Rhythmic"activity" Tensional"trajectory"

105$ 107$ 109$ 111$ 113$ 115$ 117$ 119$ 121$ 123$ 125$ 127$ 129$ 131$ 133$ 135$ 137$ 139$ 141$ 143$ 145$ 147$ 149$

Fig. 2.7. Parametric changes in the fifth section of Nympha.

starting at b.111 is the peak of an unsteady, but general increase in the dynamic
profile parameter from the beginning of the work. Also at b.111, the first violin and
cello perform ascending and descending glissandi respectively, so that the first violin
is playing repeated down-bows on A6 while the cello does the same on its low D#.
The repeating of these pitches by the first violin and cello highlights the lack of
harmonic change in this section. Up to this point, the rate of harmonic change has
been very slow, but this section contains no harmonic change. Harmonic tension also
fluctuates less in this section, again veering either side of the zero value. This could
be attributed to changes in the harmonic density parameter in contrast to much of
the preceding material, this section sees a prevalence of sustained single pitches
being played on two to three instruments.

In the context of the overall form of Nympha, the use of techniques that contrast with
what is found elsewhere in the work, such as sustained pitches, repetitive material
and the foregrounding of rhythmic or melodic material, create distinctive points of
focus that contribute to a clear formal structure. Another notable feature of this
section that gives it a strong formal function is the change in foreground
ornamentation, with the noise parameter subsiding and the rhythmic activity

! 29!
parameter coming to the foreground. The parameter of rhythmic activity comes to
the foreground in two different forms (Ex. 1.4): firstly, through the proto-melodic
material in the viola part, a fast wave-like melodic contour that also appears briefly
in the violin parts, dovetailed between the three instruments; secondly, through the
periodicity of the aforementioned sustained pitches in the violins and cello alongside
the violas proto-melodic material. Both of these rhythmic features but especially
the sustained pitches are notable for their repetition, which contrasts with the
often amorphous rhythmic context of Nympha. The scarcity of clearly repeated
material in Nympha results in this section being perceived as structurally significant,
reinforced by the expansive registral spread between the sustained pitches.

Ex. 1.4. Repetitive proto-melodic material (violins, viola) and sustained pitches (cello) in
the fifth section of Nympha.

SECTION 6 : bb.151207
While the sixth section is a continuation of the climactic function of the preceding
section, the differing tensional processes employed in each of these sections

! 30!
necessitates them being regarded separately. The tensional quality of the sixth
section is dictated more by changes to the sound/noise, harmonic density, dynamic
profile and rate of harmonic change parameters than by changes to the
inharmonicity parameter (Fig. 2.8). Though the degree of inharmonicity in this
section generally increases and fluctuates more widely in comparison with the
previous section mostly due to the increase in harmonic density it is
overshadowed by the dominance of the dynamic profile and sound-noise
parameters. After the overlapping niente lines in the violin parts at b.150 (another
feature that creates a brief but clear separation between the fifth and sixth sections),
the quartet unites in a bowed tremolo using extreme bow pressure that builds to
fortissimo (Ex. 1.5). The noise produced by the bow pressure is such that it buries the
actual pitches being fingered by the players. This united fortissimo is louder still than
the forte passage in the fifth section, continuing the sense of climax. This passage is
also the first time in the work that the surges of extreme bow pressure occur in
unison across the whole quartet, rather than being staggered across individual
instruments. As touched on in the discussion of the fourth section, this type of
change in the organisation of bow pressure surges is significant due to its creation of
a strong tensional process using noise and also due to the unity between the
instruments functioning as a point of arrival in the formal structure of the work.
Focus is also put on the unison surges of bow pressure by extending their duration
for several bars, in contrast to their lasting only a few beats when played by a single
player in previous sections.

For the first time in Nympha, the rate of harmonic change accelerates strongly.
Between bb.153189 the rate of harmonic change is at its fastest in the entire work.
An increased rate of harmonic change is of course a common tensional device in
classical music. In Nymphas atonal setting and the noise-dominated soundworld of
this section, there is no goal-oriented aspect to this increased rate of harmonic change
other than the implied sense of accelerated development, which is again reinforced
through parametric unity between the players. Throughout this section the quartet
moves together between waves of extreme bow pressure and more quiet, but
animated, arhythmic flutters of notes. At the peak rate of harmonic change the
quartet moves through a series of chords from different underlying harmonic fields,
reaching each chord via synchronised glissandi. A unique feature of this section is
the treatment of registral spread. Much of the section is restricted to a range of one-
and-a-half octaves, with the lowest pitch rarely dropping below C4. The range is
thinnest during the surges of noisy bowed tremolo, and widens to include the upper
register for the material between surges. This constriction of the registral range

! 31!

151" 153" 155" 157" 159" 161" 163" 165" 167" 169" 171" 173" 175" 177" 179" 181" 183" 185" 187" 189" 191" 193" 195" 197" 199" 201" 203" 205" 207"
Dynamic"pro5ile" Harmonic"density" Registral"spread" Tensional"trajectory"


151" 153" 155" 157" 159" 161" 163" 165" 167" 169" 171" 173" 175" 177" 179" 181" 183" 185" 187" 189" 191" 193" 195" 197" 199" 201" 203" 205" 207"
Sound/noise" Rhythmic"activity"


151$ 153$ 155$ 157$ 159$ 161$ 163$ 165$ 167$ 169$ 171$ 173$ 175$ 177$ 179$ 181$ 183$ 185$ 187$ 189$ 191$ 193$ 195$ 197$ 199$ 201$ 203$ 205$ 207$


Fig. 2.8. Parametric changes in the sixth section of Nympha.

comes immediately after the very wide registral spread of the previous section, a
clear parametric contrast between two sections that cover the first climax of Nympha.
This parametric change highlights Saariahos ability to generate tension and a sense
of formal definition through varying the parameters used to achieve this. This
section uses faster changes between states of tension and relative relaxation,
evidenced by the changes between the various fortissimo bowed tremolos and the
quieter material that lies between. Like the accelerated rate of harmonic change that
follows, these changes add to the climactic sense of momentum. The section
concludes with a high degree of rhythmic unison from bb.195207, a passage that

! 32!
Ex. 1.5. United use of extreme bow pressure in the climactic sixth section of Nympha.

sees the rhythmic activity and sound/noise parameters sharing the foreground, with
the quartet uniting in rhythmic, accented stabs of extreme bow pressure. As
mentioned earlier, this parametric interaction contrasts with the alternating fashion
in which these parameters have provided foreground ornamentation thus far, and is
the only example of this type of parametric interaction in Nympha other than in the
second climax of the work, where this interaction is drawn out further. The
combining of these two parameters creates a clear point of arrival.

SECTION 7 : bb.208243
The seventh section serves as another formal landmark due to its sustained low
tension and introduction of new timbral material, and is notable for its manipulation
of tension without relying on dynamic profile. The section begins with a haunting
change in Nymphas aural tapestry as the players whisper syllabic fragments amid
the bowed tremolos that connect with the end of the sixth section. The number of
parameters in high-tension states is reduced in this section. By the start of this section
the rate of harmonic change has dropped back to its pre-climax rate, and the
dynamic profile is never louder than mf. The registral spread opens up once again
(Fig 2.9), however the bowed tremolos cover the same wide registral range of the
fifth section, with the cello playing its open C-string for the first time in the work.

! 33!
This section is of interest for the way in which foreground ornamentation is managed
through parametric changes. Given their unpitched nature, the use of spoken voice
may be considered as a phenomenon that is quantifiable on the sound/noise
parameter, as the voices bring a new timbral dimension to the work, immediately
drawing aural attention. Along with the voices, the tensional processes at work in
this section include the aforementioned bowed tremolos, now performed in a lower
dynamic range. Surges of bow pressure have returned to their previous form of
distribution, being scattered throughout the section, played by individual players for
shorter durations. It is a moment, however, in which rhythmic material comes to the
foreground that gives this section such a unique formal character. From bb.214218 a
short motif appears in the first violin part six semiquavers alternating between D
and Eb. This proto-melodic motif is played pizzicato, a technique that up to this point
in Nympha has only appeared at one other time as a fleeting cluster in the second
violin and viola parts in b.188. The pizzicato motif is then imitated in the viola part a
major sixth lower, played arco by the first violin at its original pitches and played arco
by the second violin a minor third lower. This short imitative exchange between
members of the quartet is a clear departure from the dominant mode of discourse in
Nympha. The pizzicato technique possesses an attack profile that contrasts with that
of arco techniques, and the conspicuous response-like interaction between the parts
contrasts with Saariahos use of the string quartet in Nympha as a more globalised
sound object with little in the way of prolonged focus on a single player.


208" 210" 212" 214" 216" 218" 220" 222" 224" 226" 228" 230" 232" 234" 236" 238" 240" 242"
Dynamic"pro3ile" Rhythmic"activity" Registral"spread" Tensional"trajectory"

Fig. 2.9. Parametric changes in the seventh section of Nympha.

Following the pizzicato passage, the foreground ornamentation of the section is

again dominated by scattered bow pressure surges and whispered syllabic
fragments. It seems that the significance of this section lies in the structural
importance of the whispered text in the overall form of Nympha, rather than the
manipulation of tension. Given that it comes after the first climax of the work, this

! 34!
sections recession of the dynamic profile and rate of harmonic change parameters
suggests another rest area. Yet interest is maintained through the introduction of
new sound elements. Following the parametric point of arrival in the preceding
climactic sections, the introduction of voices and, briefly, imitative motivic material
(some of which is played pizzicato), bring renewed interest to the works surface
material. These new elements can be seen as part of some type of coda symbolising
the conclusion of one larger section of Nympha (indeed, the double barline that
separates Nympha into its two parts is not far away at b.262), with the work breaking
through to a new area that suggests a fresh timbral trajectory being established. The
imitative pizzicato passage proves to be a stylistic and timbral non-sequitur, while
the whispered voices go on to play a role in providing a sense of structural similarity
between Nymphas two halves.

SECTION 8 : bb.244262
The eighth section is made up of the remainder of Nymphas first half, which
concludes at the double barline at b.262. The most obvious formal feature of this
section is the timbral development derived from bowing on the bridge. This
technique actually enters near the end of the previous section, at b.235, but remains
through the implied separation between the two sections and is in all instruments by
the start of the eighth section. Along with this timbral development, the trajectory of
registral shape also provides clear start and end points for this section, articulating
the overall form of Nympha (Fig. 2.10).

Starting with an F6 played as an artificial harmonic by the cello, the registral spread
of this section steadily widens, retaining this upper reach while also steadily
bleeding down to include the cellos low C# by the end of the section. As mentioned
earlier, such wide registral spacing of chordal material is often applied to reinforce
cadential, climactic moments. Though we find this wide registral spread at the
conclusion of the first half of Nympha, several other parametric changes that might
be expected to accompany it are in fact subverted. The dynamic profile of this section
remains at the same general level as the previous section and gradually tapers off to
an eventual niente at the end, with the quartet playing bowed tremolos col legno
tratto, which adds some significant wispy white noise to the strings timbral

While the change in dynamic profile effectively suggests a winding down in this
section, the rate of harmonic change increases a change that, in conjunction with
the increase in registral spread, gives a contrasting impression of a climactic event

! 35!

244" 245" 246" 247" 248" 249" 250" 251" 252" 253" 254" 255" 256" 257" 258" 259" 260" 261" 262"
Dynamic"pro6ile" Harmonic"density" Registral"spread" Rhythmic"activity" Tensional"trajectory"


244$ 245$ 246$ 247$ 248$ 249$ 250$ 251$ 252$ 253$ 254$ 255$ 256$ 257$ 258$ 259$ 260$ 261$ 262$


Fig. 2.10. Parametric changes in the eighth section of Nympha.

unfolding. Indeed, other techniques employed in the latter stages of this section
suggest an active, yet muted, conclusion. The inharmonicity parameter in this section
actually reveals little harmonic tension. In fact inharmonicity levels fall into
consonant values for the majority of the section. Yet this is another one of the few
instances in Nympha where the usually closely related inharmonicity and harmonic
density parameters diverge. Double stopping is used extensively in the first half of
this section, but harmonic tension remains low. As mentioned in the methodology,
extreme registral spread has the effect of neutralising the perception of harmonic
consonance and dissonance, diminishing the effect of both. This seems to apply to
this section of Nympha, which does not exhibit any audible harmonic consonance.
Additionally, the other ubiquitous method of disguising pitch in this work the use
of constantly changing bowing positions constantly affects pitch perception,
particularly the col legno tratto playing that diminishes the perceptibility of the
pitches in the final few bars of the section.

Despite the focus here on qualities that relate to the sound/noise parameter, the
parameter that comes to occupy the foreground most clearly in this section is
rhythmic activity. Whereas the first half of this section is akin to the wash of pitches

! 36!
described in previous sections (but with the glassy overtones of the on-the-bridge
bowing providing additional timbral interest instead of surges in bow pressure), the
second half sees the quartet switch to rapid semiquaver patterns of similar contour
that shift in and out of rhythmic unison. The subdivisions then increase to
demisemiquavers at b.256, adding to the sense of tension and momentum also
reflected in the rate of harmonic change, before dissolving into bowed tremolos for
the final two bars of the section.

The final two bars of the section are significant for a rare moment of literal repetition
due to the reprisal of the cadential C# chord of the second section (Ex. 1.6). Here the
chord is initially denser than in the second section. While only the cello and viola are

used, thinning out to C#-B-D#-C, with the upper and, more importantly, lower pitch
tremolos in the second section, here all players do so. The low dynamic profile of the
chord also remains, its muted cadential quality making it recognisable as material of
some significance in the context of the work.

Ex. 1.6. Second statement of C# chord, in the eighth section of Nympha.!


! 37!
SECTION 9 : bb.263295
As the first section of Nymphas second half, the ninth section gradually accumulates
tension in preparation for the climactic section that follows it. Musical parameters are
tightly controlled in this section, with the reduction in the rate of parametric activity
underscoring the audible change of pace following the first half of the work (Fig

The use of inharmonicity as a tensional device is clearly demonstrated in this section,

as Saariaho creates surges in inharmonicity whose peaks become gradually higher.
Across the entire section, these undulations can be seen to follow an overall increase
in harmonic tension. This control of harmonic tension, which places it within a
limited bandwidth of inharmonicity over an extended duration, makes this section
unique in the context of Nympha and certainly distinguishes it as the beginning of a
new movement. Like the inharmonicity bandwidth, harmonic density remains
rather consistent throughout this section, with virtually no rests in any of the parts.
The dynamic profile of this section also follows a similar trajectory to the
inharmonicity parameter, rising and falling but maintaining a steady overall increase
in its slow crescendo towards the next section. This is achieved to some extent
through the periodic dynamic surges in the players parts that, as with much of the
first half of Nympha, are not synchronised. Over the course of this section, however,
synchronisation is again used as a device for gathering force and momentum. The
players dynamic surges eventually begin to synchronise, and as the surges in the
individual parts are also growing steadily louder, the overall dynamic profile
reaches gradually higher peaks as the surges become more unified.

As with the parameters already discussed in this section, registral shape also remains
generally consistent, with subtle changes unfolding over the duration of the section.
Predominantly occupying the middle and low registers, this section also contains
frequent flashes of artificial harmonics more than two octaves above the treble staff.
As the dynamic surges become more synchronised, the registral spread grows
through the use of artificial harmonics pushing further above the treble staff. The
rate of harmonic change is also steady in this section, and overall slightly faster than
the rate of harmonic change observed in the non-climactic sections of Nymphas first
half. Between bb.287295, the three dynamic surges which, by this stage in the
section, are closely synchronised between the players all occur in conjunction with
a sharp increase in registral spread that then decreases in tandem with the dynamic
decrescendos. The three spikes in dynamic profile are of course more pronounced

! 38!

263" 265" 267" 269" 271" 273" 275" 277" 279" 281" 283" 285" 287" 289" 291" 293" 295"
Dynamic"pro5ile" Harmonic"density" Registral"spread" Tensional"trajectory"


263$ 265$ 267$ 269$ 271$ 273$ 275$ 277$ 279$ 281$ 283$ 285$ 287$ 289$ 291$ 293$ 295$


Fig. 2.11. Parametric changes in the ninth section of Nympha.

due to the closer synchronisation between the individual players. The final five bars
of the section see its dominant playing style disturbed in preparation for the
climactic section that follows, as in b.291 each players crescendo coincides with a
pitch that is played dtach, with the duration of notes also becoming longer and the
vibrato instructions having disappeared. This is the first time such a clear tensional
relationship between the dynamic profile and registral spread parameters is used in
Nympha, highlighting Saariahos heavy use of parametric desynchronisation
elsewhere in Nympha. This parametric desyncronisation is a tensional device used
for subverting potentially climactic passages, which in turn reinforces the tensional
impact of true climactic passages in the work, where most parameters are

Another obvious feature of this ninth section is the absence of foreground material
based on changes to the sound/noise and rhythmic activity parameters, which
quickly sets apart the two halves of Nympha. As an exception to what has preceded
it, this section can be perceived as a point of contrast within the context of Nympha,
marking a new approach to the foreground-background arrangement of musical
elements in the wake of the first half of the work. Rhythmic activity is minimal

! 39!
throughout this section, with pitches flowing in a slow, arhythmic manner akin to
the wash of pitches passages described in the first half of Nympha.

Without the surges of bow pressure and other variations of articulation lifting a
single player out of the overall sound in the first half of the work, the ninth section
sees the quartet organised as a more globalised texture. Interestingly, this is achieved
through the works most contrapuntal material. As a slower realisation of the wash
of pitches described in earlier sections, here the registral extremes of high artificial
harmonics or low cello pitches are the clearest features that remind the listener of the
individual parts that comprise the ensemble. The quartet interacts in its most
traditionally contrapuntal organisation in this section, yet it is articulated in such a
way that the typical clarity of individual lines is often obscured by its languishing
rhythmic character.

The timbral backdrop of this section retains a degree of familiarity due to the
continuation of changes in bow position, whose movements between sul tasto and sul
pont also gradually synchronise with the dynamic surges. In a departure from the
way bow position is prescribed in the first half of Nympha, peaks in dynamic profile
now coincide with reaching sul pont. In the first half of Nympha, peaks in dynamic
profile coincide with naturale bow position. This contrary relationship between
bowing position and dynamic profile, in which purity of pitch is compromised at the
moment it reaches a dynamic peak, illustrates yet again the emphasis on
desynchronisation and subversion of energetic profile in Nympha. Adding to this
effect is a notable instruction in this section regarding vibrato that is overlying the
bow position instructions. Like the bow position instructions, the vibrato instructions
continually move between vibrato and senza vibrato, with peaks in dynamic profile
coinciding with reaching senza vibrato. Here the decoupling of articulation and
dynamics subverts their traditional relationship, frustrating the expressive capability
of the material.

SECTION 10 : bb.296334
The tenth section delivers the climax threatened in the ominous ninth section by
employing material and tensional processes that contrast heavily with the ninth
section. This section reprises many of the processes observed in the climactic material
of the sixth section, pushing those processes to their extremes while also creating a
structurally coherent link with the first half of Nympha (Fig. 2.12). Like the climactic
material in the first half of the work, this section serves as a point of arrival once
again in its use of parametric synchronisation, employing increased synchronisation

! 40!

296" 298" 300" 302" 304" 306" 308" 310" 312" 314" 316" 318" 320" 322" 324" 326" 328" 330" 332" 334"
Dynamic"pro4ile" Harmonic"density" Registral"spread" Tensional"trajectory"


296" 298" 300" 302" 304" 306" 308" 310" 312" 314" 316" 318" 320" 322" 324" 326" 328" 330" 332" 334"
Sound/noise" Rhythmic"activity"

Fig. 2.12. Parametric changes in the tenth section of Nympha.

between the players to bring this climax into focus, while retaining a degree of flux
as it jumps between a number of techniques and motifs (Ex. 1.7). Between bb.296325
the quartet moves together briskly between 1) surges of extreme bow pressure
played in polyrhythmic semiquaver and demisemiquaver glissandi patterns; 2) a
version of the animated, arhythmic flutters of notes found in the climax of the sixth
section, now transformed by heavily accented and staccatissimo playing; and 3)
downward glissandi with extreme vibrato.

Though this section occupies a state of high tension due to its explosive rate of flux,
in a formal sense it can also be considered a long-awaited release from the tension of
the tightly controlled and consistent material of the ninth section. As with the sixth
section, the tensional quality of the tenth section is dictated more by changes to the
sound/noise, harmonic density, dynamic profile and rate of harmonic change
parameters than by changes to the inharmonicity parameter. It is not possible to
accurately measure inharmonicity in this section due to the extensive use of
glissandi. At many points throughout this section, at least one player is performing a
glissando technique, meaning that chordal material is not stable enough to enable
extraction of distinct pitches for analysis. It is not within the scope of this analysis to

! 41!
Ex. 1.7. Flux in articulations and techniques in the climactic tenth section of Nympha

develop a system for the measurement of inharmonicity of chordal material

containing glissandi. It is possible, however, to acknowledge the pitch instability that
results from glissandi techniques and their effectiveness as a pitch-based tensional
process. Therefore, it is appropriate to shift focus in this section from inharmonicity
to harmonic instability, which better describes the microtonal nature of the glissandi
transitions between pitches. Harmonic instability still implies a tensional process
concerning the degree to which a pitch, equal-tempered or not, remains static for its
duration. This temporary reframing of inharmonicity obviously affects the accuracy
with which the rate of harmonic change can be measured in this section as well,
though there is enough harmonic material without glissandi to still allow fairly
accurate identification of changes in the underlying harmonic field.

Having identified these obstacles to obtaining accurate measurement, it must be

noted that a significant amount of the harmonic material in this section is ultimately
overshadowed, as it is in the sixth section, by the dominance of noise. Periodic surges
of extreme bow pressure return in this section, performed in unison by the entire
quartet. Whereas in the sixth section these surges often contain glissandi, in the tenth
section the surges are always combined with glissandi, in every part. All surges from
bb.296324 are also played sul pont, resulting in a combination of pitch instability and
timbral instability combined with the most immediate and foregrounded parametric
extreme noise which straddles both pitch and timbral instability.

The rate of harmonic change is accelerated in this section, though not quite to the
extent of the climactic material in the first half of the work, but drops away
significantly for Nymphas tensional apex starting at b.326. The passage that extends

! 42!
from b.326334 is so dominated by noise that the rate of harmonic change is
essentially irrelevant here. The tenth section shares with the sixth section the loudest
dynamic profile of the whole work, with the surges of bow pressure being extended
to their longest durations. From bb.326334, the workss main climax a hocketed
pulse of ferocious, sustained down-bows (Ex. 1.8) breaks free from the
polyrhythmic tension and fast changes of material that have dominated thus far. This
hocketed, down-bowed passage marks the strongest fusion of the two main
foreground ornamentation parameters in Nympha, sound/noise and rhythmic
activity. While the dominance of noise in this passage is such that the inharmonicity
and rate of harmonic change parameters are rendered irrelevant, the other
parameters dynamic profile, harmonic density and registral spread are
employed in ways that typically serve to maximise the impact of climactic material.
The extremely high dynamic profile and wide registral spread (between three and!

Ex. 1.8. Hocketed pulse of extreme bow pressure in the climactic tenth section of Nympha

! 43!
four octaves) combine with the foregrounded noise and rhythm to finally realise a
climax that combines! maximised timbral, rhythmic and dynamic parameters (with
the noise parameter so extreme that it effectively consumes the inharmonicity and
rate of harmonic change parameters), underlining the singular significance of this
passage in the context of Nympha. The rhythmic parameter is adjusted between
bb.328334 to gradually release tension in this passage and eventually reach a point
of rest. The durations of the sustained down-bows become increasingly extended
and the dynamic profile decreases, and eventually the sense of rhythmic pulse is lost.
Changes to slower tempi are also employed to this end.!

SECTION 11 : bb.335362
The eleventh section continues the low-tension trajectory unfolding by the end of the
preceding section. Several parameters are reorganised for this section, however, so
that the low degree of tension is generated in a different way. We see the foreground
ornamentation of this section (Fig. 2.13) dominated by the rhythmic activity of the
proto-melodic material in the first violin, and a return to a more consistent use of
material, as in the ninth section. The eleventh section begins with the first violin
performing a repeated ascending figure, played legato. The bow position for the first
violins line is consistently naturale, a notably stable departure from the restless
bowing instructions in the rest of the work. While the rhythm of the proto-melodic
material differs slightly with every iteration, its general range and underlying
harmonic field remains quite consistent. The sound/noise parameter is not
foregrounded despite the rest of the quartet playing slowly rising and falling
glissandi figures, their col legno tratto bowing and piano dynamic relegates this
material to a glassy atmospheric backdrop for the first violins timbrally pure
material. The first violins proto-melodic material from bb.335346 appears to be an
extended recasting of the second violins proto-melodic material that emerged in the


335" 336" 337" 338" 339" 340" 341" 342" 343" 344" 345" 346" 347" 348" 349" 350" 351" 352" 353" 354" 355" 356" 357" 358" 359" 360" 361" 362"
Dynamic"pro6ile" Registral"spread" Rhythmic"activity" Tensional"trajectory"

Fig. 2.13. Parametric changes in the eleventh section of Nympha.

! 44!
opening section of Nympha. As with the climactic content of the previous section, the
return to material from an earlier section here reinforces a structural coherence that
links the two halves of the work. Again, the presence of glissandi for much of this
section imposes an obstacle for measuring harmonic tension and rate of harmonic
change. Since the bowing techniques applied to all players excluding the first violin
up until b.348 effectively mask their notated pitches, it is only the first violin that can
be drawn upon for accurate information about the rate of harmonic change. Reliable
measurement of the rate of harmonic change becomes possible from b.354, once
glissandi no longer appear in any parts. The first half of this section, containing the
first violins proto-melodic material, displays very little harmonic change, reinforcing
the sense of stasis created by the repetitive nature of the first violins ascending lines.
The rate of harmonic change accelerates slightly once the glissando material is
replaced by more stable, pure pitches, but not enough to generate any sense of
momentum. The dynamic profile of this section never rises beyond mp, which is
assigned predominantly to the first violin only.

The final five bars of this section (Ex. 1.9) recall the muted cadential moments at the
conclusion of other sections of Nympha bb.3137 (end of the second section) and
bb.261262 (end of the eighth section and first half of Nympha). Again, this material

contains a low dynamic profile, wide registral spread and similar pitch material (C#-

Ex. 1.9. Third statement of C# chord, in the eleventh section of Nympha.!

! 45!
G-A), with the most identifiable and important pitch of course being the cellos low
C#, played tremolo. Though the harmonic density of the chords based on the low C#
differs with each occurrence, it can be reliably inferred that these pitches are derived
from a common underlying harmonic field. When considering that in addition to the

ubiquitous tremolo technique, the chords based on the low C# also share common
dynamic and registral parametric features, it appears that Saariaho is using
recognisable repetition of this harmonic material at structurally important points
(early in the work, at the conclusion of its first half and near the end of the second
half) to tie the work together through quasi-cadential points of arrival that reveal
Nymphas home key, or home harmonic field.
SECTION 12 : bb.363389
The final section of Nympha continues the decline in energy towards silence, while
reprising particular types of foreground ornamentation that illuminate formal
similarities between the two halves of the work (Fig. 2.14). The section begins with
the return of whispered material, as used near the end of the first half of the work.
The only variation in this second appearance is that the syllabic fragments are now
full lines of text, taken from But there has to be more by Russian poet Arseniy
Tarkovsky. Lines are whispered in full, though some syllables are drawn out across
several beats. The syllables that are drawn out are also the same syllabic fragments
used in the first half of Nympha, their source now revealed.

The use of the whispered text near the conclusion of both halves of Nympha adds a
further structural landmark and creates a sense of formal logic for the work. As
mentioned in regard to the seventh section, the whispered text has value as a timbral
refresher that draws the attention of the listener. This also applies in the second half
of the work, with the whispered text following sections that have pushed the
extremes of the sound/noise parameter to the foreground for an extended period
(tenth section), with rhythmic activity also featuring.

Rhythmic activity occupies the foreground of the early part of the twelfth section,
alongside the whispered text, with the cello playing a passage of rapid but crisp
microtonal contours initially covering less than an octave. Here we have another
example of climactic material subverted the cello line, like the whispered text and
the accompanying upper register pitches held by the violins, are played at low
volume. The combination of low and high pitch material with nothing occupying

! 46!

363" 364" 365" 366" 367" 368" 369" 370" 371" 372" 373" 374" 375" 376" 377" 378" 379" 380" 381" 382" 383" 384" 385" 386" 387" 388" 389"
Dynamic"pro6ile" Harmonic"density" Registral"spread" Tensional"trajectory"


363" 364" 365" 366" 367" 368" 369" 370" 371" 372" 373" 374" 375" 376" 377" 378" 379" 380" 381" 382" 383" 384" 385" 386" 387" 388" 389"
Sound/noise" Rhythmic"activity"

Fig. 2.14. Parametric changes in the twelfth section of Nympha.

the treble register of this spread generates tension, but this can be considered
registral tension as opposed to harmonic tension, as the great distance between the
low and high pitches lessen the audible intervallic relationships that we interpret as
consonant or dissonant. The fast playing in the cello line also conveys a sense of
momentum, but the effect is subverted by the quiet dynamics and the avoidance of
clear melodic or tonal direction in the line. Additionally, the rate of harmonic change
remains slow throughout this section, becoming gradually slower following the
conclusion of the whispered text passage at b.380.

Also decreasing in this section is harmonic density, which remains relatively low
throughout, in contrast to the thicker harmonic material that dominates the second
half of Nympha. This reduction in harmonic density is another means of reducing
tension as the work winds down. As seen throughout the work, and especially in the
first half, the sound/noise and rhythmic activity parameters often alternate in their
role of providing surface material in Nympha. This is seen again in the twelfth
section, once the whispered text and cello line conclude. The sound/noise parameter
now comes to the foreground for the final few bars of Nympha, with two sustained

! 47!
surges of extreme bow pressure played without tremolos by the cello. The rest of the
quartet accompanies the cello with high artificial harmonics, again using registral
tension of wide harmonic spacing in addition to harmonic tension. The degree of
inharmonicity is rendered unstable due to the quartet playing trills. The effect of the
trills is reminiscent of the glissandi discussed in the tenth section, which also
generates tension through harmonic instability. The closing bars of Nympha suspend
this harmonic tension and instability, while the other parameters descend to their
zero values. Eventually the cello is left alone, avoiding release to the end with a trill
flickering away into silence.

In summary, Saariaho creates a coherent, tension-based formal discourse in Nympha

by relying on a core group of parametric relationships and combinations, while also
using non-literal repetition of selected material. Synchronisation of rhythmic
activity, dynamics or noise is key to any one parameter dominating the material,
its importance coming from its relative scarcity in the overall context of the work.
Noise and rhythmic activity are clearly the parameters that Saariaho uses to
construct foreground ornamentation, while traditionally tension-related parameters
such as inharmoncity and dynamic profile also feature. The dominance of noise in
climactic passages is such that it obscures the perception of an accelerated rate of
harmonic change, Saariaho instead pairing noise with rhythmic activity to create the
rhythmicised noise that heightens Nymphas climactic passages.

! 48!

Du Cristal
for orchestra

Where Nympha relies on changes in the foregrounded sound/noise parameter for

generating tension and release, Du cristal primarily uses the dynamic profile and
inharmonicity parameters to achieve this. Additionally, where foreground
ornamentation in Nympha is provided through foregrounded activity in the
rhythmic and sound/noise parameters, Du cristal relies purely on rhythmic activity
to provide microstructural interest as the work unfolds. Rhythmic activity is
manipulated in two main ways to reinforce formal structure in Du cristal. Firstly, it is
used in a similar fashion to Nympha, with rhythmic unison being an important
device for creating points of arrival in the work. The orchestra occasionally aligns
with a perceivable rhythmic grid, changing the works relationship to time by
generating brief temporal momentum. Secondly, rhythmic activity is used as a
device for creating passages that change the listeners perception of the orchestra in
this work as a global sound object, turning attention to the numerous pockets of
individual activity and contrapuntal lines that can populate its thick texture.
Saariaho uses multiple layers of polyrhythmic material to create activity that is
nested within dense orchestral textures, varying the degree to which this
polyrhythmic material comes to the foreground. In these instances it is impossible to
perceive an underlying rhythmic grid, but attention is turned to the individual lines
that emerge from the sea of sustained pitches that comprise a timbre chord.
SECTION 1 : bb.136
The first section of Du cristal displays the slow parametric changes that typify much
of the work (Fig. 3.3). It consists of one harmonic field, with certain combinations of
pitches from within this harmonic field being illuminated at different times, as
described by Hargreaves (2011). Harmonic tension is manipulated through variation
of the pitch combinations.

As with the rate of harmonic change, the changes in dynamic profile in this section
are also slow. However, the early moments of the work are notable for their loud
dynamics opening the work at a fortissimo dynamic provides a strong and obvious!
parametric contrast to the silence from which it originates, but the intent behind this
loud dynamic is that it serves as a starting point for the trajectory of the dynamic!

! 49!
Section Bars Time Tensional Main features

1 136 0:001:51 Transition out of sustained,

high to low tension
globalized sound-mass with
introduction of short
arhythmic phrases for some

2 3764 1:522:59 low tension, crescendo rest area, gradual increase in

at end dynamic profile, pitch
instability techniques used for

3 65107 3:005:16 rest area, larger Another gradual increase in

crescendo at end dynamic profile, very wide
registral spread, long absence
of any foregrounded
ornamentation, concludes with
parallel parametric changes
towards high tension

4 108140 5:177:04 High dynamic profile,

high tension
harmonic density and registral

5 141244 7:0510:31 high to low tension Rhythmic activity parameter

as a tensional device
(percussion solo), infrequent
harmonic tension, high
dynamic profile

6 245277 10:3211:34 low to high tension Rise in harmonic tension,

though climax uses reduction
of registral spread to one pitch

7 278329 11:3514:11 low to high tension More synchronized

relationship between increases
in dynamic profile and
density, use of pitch instability
to add tension ahead of
following section

8 330353 14:1215:31 high tension to low Climactic passage using

tension rhythmic activity, and
dynamic profile. Low
inharmonicity but extreme
registral spread.

9 354368 15:3216:40 low tension Focus on percussion

synchronization. Extreme
registral spread but low
dynamic profile.

Fig. 3.1. Formal outline of Du cristal

! 50!





0:00" 1:00" 2:00" 3:00" 4:00" 5:00" 6:00" 7:00" 8:00" 9:00" 10:00" 11:00" 12:00" 13:00" 14:00" 15:00" 16:00"


0:00" 1:00" 2:00" 3:00" 4:00" 5:00" 6:00" 7:00" 8:00" 9:00" 10:00" 11:00" 12:00" 13:00" 14:00" 15:00" 16:00"

Fig. 3.2. Rate of harmonic change and tensional trajectory in Du cristal

profile parameter towards the p dynamic reached at the end of the section. The
overall form of this section is a gradual shift from high tension to low tension, with
the trajectory in dynamic profile being the most crucial parametric change in creating
this shift.
Additionally, rhythmic activity is utilized with differing relationships to the
underlying meter to assist this sections tensional transition. Firstly, a modest

10$ !10$

0$ 0$
1$ 2$ 3$ 4$ 5$ 6$ 7$ 8$ 9$ 10$ 11$ 12$ 13$ 14$ 15$ 16$ 17$ 18$ 19$ 20$ 21$ 22$ 23$ 24$ 25$ 26$ 27$ 28$ 29$ 30$ 31$ 32$ 33$ 34$ 35$ 36$

Dynamic$pro7ile$ Harmonic$density$ Registral$spread$ Tensional$trajectory$ Inharmonicity$(harmonic$tension)$

Fig. 3.3. Parametric changes in the first section of Du cristal.

! 51!
amount of rhythmic momentum is generated between bb.119 through an F6 played
in repeated semiquavers passed from glockenspiel to piccolo and clarinet (though
initially introduced by the unpitched triangle). The high registral placement of this
pattern precludes the domination of any rhythmic drive but does support an overall
state of forward movement. As the section progresses, rhythmic activity is utilized to
transition out of the sustained, globalized sound-mass of the opening timbre chords,
with short arhythmic phrases or trills from woodwind and string instruments being
highlighted momentarily, beginning to pull attention away from the textural to the
contrapuntal, as individual phrases gradually overlap and dovetail. As with the
repeated semiquavers that preceded them, these flickers of activity from individual
instruments provide the foreground ornamentation in this section, yet the
cumulative effect of the freer playing is reminiscent of the arhythmic quality found
in Nympha. The general formal trajectory of this section can be described as moving
from a high-tension global sound-mass toward a medium-tension state in which
increased individual rhythmic activity draws more attention to individual lines
within the orchestral texture.

It is notable that while registral spread remains static until b.23, harmonic density
has increased by b.17, filling in the chordal material with additional pitches. As
found in Nympha, the increase in harmonic density leads to a parallel increase in
inharmonicity. In Du cristal, harmonic tension steadily rises from the opening of the
work through to b.17, with the fortissimo dynamic remaining in place until around
this point also. From bb.1736, the tension of this section begins to unravel. The
harmonic density and inharmonicity parameters drop, stabilising at b.26 with lower
values than those found at the opening of the work. Meanwhile, the gradual drop in
dynamic profile starting at b.13 contributes to the overall tensional transition of the
section. The already wide and static registral spread begins to fluctuate slightly from
b.23 and becomes wider still, most notably through the introduction of the low C on
the piano. The increase in registral spread combined with the gradual decrease in
harmonic density makes for wider spacing between pitches, particularly in the
middle register, which enhances the contrasting spaciousness of this half of the
section with the sustained density of the first half.

SECTION 2 : bb.3764
The second section is an extension of the tensional state reached at the end of the first
section and largely serves as a rest area, similar in function to those described in the
analysis of Nympha. While harmonic density and inharmonicity are less than that of

! 52!
the first section and fluctuate by small amounts, the underlying harmonic field
undergoes a gradual transformation in preparation for a new harmonic field in the
following section (Fig. 3.4).

Foreground ornamentation in this section is initially supplied by short overlapping

phrases in the woodwind instruments upper registers, the variety of metric
subdivisions in each part again producing an arhythmic mesh of pitches. The
woodwind phrases thin out, with the pitched percussion, harp, piano and
synthesizer entering at b.52 in a more unified motivic statement (the harp, piano and
synthesizer notably playing in unison). The piano and harp are brought out of the
contrabass register for use in this motivic statement, with the subsequent constriction
of registral spread to three and a half octaves the smallest registral spread in the
work thus far setting the section up for a gradual re-expansion of registral spread,
as the section builds to a crescendo in anticipation of the third section.

Another parameter assisting the gradual increase in tension at the end of the second
section is dynamic profile. The dynamic profile of this section follows a clear
trajectory from p at b.37 to fff at b.64. From b.55 the first change in the underlying
harmonic field becomes detectable. A degree of harmonic flux occurs as slow
glissandi are introduced to some string parts, travelling toward pitches found in the
underlying harmonic field in the following section. The pitch instability of the
glissandi enhances the tensional quality of the crescendo. bb.6264 contain an
accumulation of extreme bow pressure in the now extensively divided string section,
which dominates the orchestration at this point. While the noisier extremes of the
sound/noise continuum (achieved through extreme bow pressure) strongly intensify
the tensional state created within Nymphas smaller ensemble, the non-tremolo bow

10$ !10$

0$ 0$
37$ 38$ 39$ 40$ 41$ 42$ 43$ 44$ 45$ 46$ 47$ 48$ 49$ 50$ 51$ 52$ 53$ 54$ 55$ 56$ 57$ 58$ 59$ 60$ 61$ 62$ 63$ 64$
Dynamic$pro7ile$ Harmonic$density$ Registral$spread$
Tensional$trajectory$ Inharmonicity$(harmonic$tension)$
Fig. 3.4. Parametric changes in the second section of Du cristal

! 53!
pressure employed at this point in Du cristal does not cut through the other strings
played ord, the glockenspiel, the horn and the rolls on the bass drum and timpani.
The noise of the bow pressure seems to serve more as a device for supporting the
unpitched timbral quality of the bass drum and, to a lesser degree, the timpani, in the
heightened tensional state at the end of this section.

SECTION 3 : bb.65107
The tensional profile of the third section bears some similarity to that of the second
section a trajectory from low to high tension with the main difference being
that certain parameters are pushed to further extremes, providing this section with a
clear identity in the context of the work (Fig. 3.5). Dynamic profile is again used as a
primary driver of the tensional trajectory in this section, with other parameters
contributing as the work builds towards the climactic passage in the fourth section.

10$ !10$

0$ 0$
65$ 67$ 69$ 71$ 73$ 75$ 77$ 79$ 81$ 83$ 85$ 87$ 89$ 91$ 93$ 95$ 97$ 99$ 101$ 103$ 105$ 107$

Dynamic$pro5ile$ Harmonic$density$ Registral$spread$ Tensional$trajectory$ Inharmonicity$(harmonic$tension)$

Fig. 3.5. Parametric changes in the third section of Du cristal

A prolonged period of release opens the third section, as the crescendo of the
preceding section dissolves. Dynamics are reduced to p, with the extensively divided
string section (accompanied by the synthesizer) remaining in place. As also observed
in Nympha, Saariaho employs a somewhat subversive combination of parametric
extremes. Extremely wide registral spread and high harmonic density two
parametric extremes traditionally associated with climactic material are combined
with quiet dynamics and a minimal rate of harmonic change. The result is a period of
stasis, even in the context of the gradated parametric development found in Du
cristal. On closer inspection, until b.75 the registral spread of this section is heavily
weighted towards the upper reaches of the string instruments registers, reaching
more than two octaves above the treble staff. The remaining pitches are widely
spread, extending beneath the bass staff. There is a small amount of noise in this
passage due to the timbral quality of pitches in the upper registers of the violins and

! 54!
violas, where tonal purity becomes increasingly thin and glassy. This registral skew
and associated timbral effect is the primary source of any tension in this passage.
There is also a slight increase in inharmonicity in this passage, though, as discussed
in the analysis of Nympha, the voicing of chords at registral extremes diminished the
intensity of consonance and dissonance, another example of parametric subversion
in Saariahos writing.

As already mentioned, the parameter that most clearly reinforces the low tensional
state of this passage is the low dynamic range, which remains in place until b.78.
Though Du cristal is less consistent than Nympha in its employment of foreground
ornamentation, it is worth mentioning that the degree of stasis generated in this
passage is also reliant on the absence of any foregrounded material. Like the ninth
section of Nympha, this absence of foregrounded material in this passage of Du
cristals third section assists in creating a clear structural landmark, given its contrast
to the material that has preceded it.

Foreground ornamentation reappears at b.78, coinciding with the onset of other

underlying parametric changes as the third section starts to accumulate tension
leading into the fourth section. Glockenspiel, vibraphone, crotales and synthesizer
draw the listener in, returning with high-register polyrhythmic/arhythmic material
before a trilling piccolo line is added (b.88), along with harp and piano lines that
mimic the rhythmic character of the synthesizer (b.94). Harmonic density decreases
as the foreground ornamentation returns, and starts to gradually increase for the
remainder of the section. Inharmonicity also decreases at the same point, but remains
low for a longer period, allowing for a more dramatic increase occurring from b.97.
By this point, the registral spread has filled out, with more pitches populating the
middle and low registers, making the increase in harmonic tension more noticeable
due to the closer registral proximity of the pitches. Also occurring at this point in the
section is the beginning of a change to a new harmonic field, adding a sense of
harmonic flux to the ongoing rise in tension. The transition to the new harmonic field
takes place across the next 15 or so bars, straddling the boundary between the third
and fourth sections marked by peaking swells on the bass drum, timpani and tam-
tam at b.108. From b.96 until the end of the section Saariaho revisits the use of
extreme bow pressure in the string section, though again the noise generated by this
technique is embedded in the background as it is overpowered by the more cleanly
played material in the other sections of the orchestra. The conclusion of this section
displays a clear convergence of parameteric changes that move toward higher states
of tension, marking the first parametric point of arrival in Du cristal.

! 55!
SECTION 4 : bb.108140
The fourth section marks the beginning of a longer passage of high-tension, climactic
material in Du cristal. Building on the tensional trajectory of the previous section, the
fourth section sees some parameters reach new peaks, with tension maintained in
preparation for further intensification of tensional energy in the fifth section (Fig.
10$ !10$

0$ 0$
108$ 110$ 112$ 114$ 116$ 118$ 120$ 122$ 124$ 126$ 128$ 130$ 132$ 134$ 136$ 138$ 140$

Dynamic$pro4ile$ Harmonic$density$ Registral$spread$ Tensional$trajectory$ Inharmonicity$(harmonic$tension)$

Fig. 3.6. Parametric changes in the fourth section of Du cristal

The opening moments of the section see harmonic density reach its highest peak thus
far as the entire orchestra is active by this point, filling out thick chords. Density rises
and falls with periodic surging chords in the brass choir that act as foreground
ornamentation at the outset of this section. The increase in harmonic density
coincides with an expansion of registral spread to almost seven octaves the widest
spread so far in the work, surpassed by only a couple of later passages. As
mentioned previously, the transition to the fourth section coincides with a transition
in the underlying harmonic field, which is complete by around b.113. Inharmonicity
rises and falls along with harmonic density, often increasing past the previous peaks
of inharmonicity reached during the decay of Du cristals opening chord. Again,
inharmonicity and dynamic profile are the two main parameters generating tension
in this section. The consistent ffff dynamic range makes this section the loudest since
the opening dozen bars of Du cristal. It is the chordal surges from the brass choir that
take the role of foreground ornamentation in this section, despite the eventual
sprinkling of high-pitched percussion around these surges. The attack profile of the
brass surges transform during this section, from swift crescendos to sudden attacks
at forte or louder, an attack profile not employed since the works opening series of
chords. In combination with the high degree of harmonic tension, the sudden attack
profile of the brass chords jumping out from the backdrop of high string harmonics

! 56!
and trilling woodwinds maintains the high-tension state of this section in
anticipation of the continued escalation of tensional energy with its notable
foregrounding of rhythmic activity in the fifth section.

SECTION 5 : bb.141 244

The fifth section includes the peak of the longer climactic passage that has been
unfolding over the previous section, with the foregrounding of the rhythmic activity
parameter as a tensional device, replacing inharmonicity alongside dynamic profile
(Fig. 3.7).

Dynamic profile reaches its highest peak of the entire work in the fifth section, with
timpani (later joined by roto-toms) featured in a soloistic manner for much of the
section, often playing at ff or louder, with each percussive phrase separated by
chordal bursts from the orchestra. Though the timpanis material is fairly consistent
in its rhythmic character, the constant triplet, quintuplet and septuplet subdivisions
obstruct the perception of any underlying metric pulse, generating a high amount of
rhythmic tension. While rhythmic activity and dynamic profile are foregrounded in
this section, inharmonicity and harmonic density decrease markedly. The vast
majority of the section contains a minimal amount of harmonic tension the lowest
amount in the work thus far. It could be argued, however, that wide registral spread
comes into play here with regards to perception of consonance/dissonance, as
discussed earlier, increasing the perception of inharmonicity due to the distance
between pitches compromising the ability to perceive consonant intervallic
relationships within chordal material.

Registral spread is at its most unstable thus far, with both upper and lower registral
extremes quickly varying between one and two octaves between bb.141201. Prior to
this, the registral extremes in Du cristal had remained rather stable, in keeping with
the prevalence of sustained pitches that dominate the textural fabric of the work. The
instability of registral spread here contributes to the general sense of flux generated
by the timpanis arhythmic phrases. Extreme bow pressure is employed at several
dynamic peaks in this section, though again it does not cut through the other sections
of the orchestra, serving instead as a timbral backdrop as opposed to a foregrounded
source of tension as seen in Nympha.

The start of the fifth section is also marked by a change in the underlying harmonic
field. Again, the transition between harmonic fields is gradual, stretching across a

! 57!

141" 145" 149" 153" 157" 161" 165" 169" 173" 177" 181" 185" 189" 193" 197" 201" 205" 209" 213" 217" 221" 225" 229" 233" 237" 241"

Dynamic"pro6ile" Rhythmic"activity"
10$ !10$

0$ 0$
141$ 145$ 149$ 153$ 157$ 161$ 165$ 169$ 173$ 177$ 181$ 185$ 189$ 193$ 197$ 201$ 205$ 209$ 213$ 217$ 221$ 225$ 229$ 233$ 237$ 241$

Harmonic$density$ Registral$spread$ Inharmonicity$(harmonic$tension)$

Fig. 3.7. Parametric changes in the fifth section of Du cristal

dozen bars, including the final bars of the fourth section. A second harmonic change
takes place in the fifth section, between bb.191199, by which time the timpani/roto-
tom phrases have temporarily ceased and a brief, low-tension rest period (bb.183
204) has ensued. The dynamic profile decreases, with foreground ornamentation
being occupied by a brief return of polyrhythmic phrases from the harp, piano and
synthesizer similar to that described in the third section. Ascending, overlapping
polyrhythmic phrases then populate the woodwind choir from b.196, gradually
reinstating rhythmic instability as the section builds to a crescendo ahead of the
timpani and roto-toms resumption at b.205. Throughout the section, harmonic
density steadily increases, but inharmonicity generally remains at the same low level.

At b.227 Du cristal arrives at one of its few passages of sharper rhythmic focus, a
clear formal landmark for the work. The timpani and roto-toms depart and the
orchestra unites in a continuous pattern of semiquavers, with each instrument
repeating the same note or notes. Even in this moment of rhythmic unity, Saariaho
ensures that its edges remain blurred by varying the articulations of these notes
across the orchestra, using accented, tenuto to staccatissimo markings. Between
bb.227243 this metric grid of semiquavers propels the work along, with varying

! 58!
combinations of instruments taking up the rhythmic pattern while others
temporarily drop out or hold brief sustained notes. Here we see Saariaho once again
using synchronisation to create a unique point of arrival in a work in this instance,
using rhythmic unity at the end of an extended climactic passage containing
rhythmic activity that, up to this point, has operated without a strong sense of
temporal momentum by not working around a clear pulse.

Throughout this semiquaver passage, registral spread expands into the bass register,
so that by the end of the section the material covers six octaves. In contrast, the
dynamic profile begins to drop back from b.236, reaching mp/mf by the end of the
section. By this point, the semiquaver pattern has lost much of its tensional impact
due to the ritardando starting at b.241 and the contamination of the semiquaver
pattern with overlaid sextuplet subdivisions.

SECTION 6 : bb.245277
As also seen in the third section, the sixth section follows a general trajectory from
low to high tension, with dynamic profile again used as a primary driver of the
tensional trajectory. Inharmonicity returns as an influential tensional parameter, with
this section being unique in its more active variation in the degree to which each of
these parameters influences the tensional process. This section also uses the
minimisation of registral spread as a component in the generation of tension (Fig.

The transition to the sixth section is marked by a change in harmonic field and a
sharp rise in harmonic tension an interesting contrast to the very low amount of
inharmonicity in the climactic, high-tension fifth section. The sustained and
repetitive quality of the pitch material means that inharmonicity remains at a
constant level until b.260, when harmonic density begins to steadily decrease. By the
end of the section, harmonic density has been reduced to one unison pitch Ab4 in
the treble staff played by most of the orchestras pitched instruments. This unison
pitch is brought to one of the strongest crescendi in the entire work, made all the
more significant as a structural arrival point due to its strong contrast with the dense
harmonic context that dominates Du cristal. Aside from the novelty of this brief
sparseness in density, this section is also notable for the expansiveness and stability
of its registral spread up until its final bars. Registral spread reaches seven octaves,
the widest spread of the entire work. Much like the opening section, the registral
extremities of much of the sixth section remain static, with pitch movement taking
place within these registral boundaries. An example of this embedded pitch

! 59!

245" 247" 249" 251" 253" 255" 257" 259" 261" 263" 265" 267" 269" 271" 273" 275" 277"

Dynamic"pro5ile" Rhythmic"activity" Tensional"trajectory"

10$ !10$

0$ 0$
245$ 247$ 249$ 251$ 253$ 255$ 257$ 259$ 261$ 263$ 265$ 267$ 269$ 271$ 273$ 275$ 277$

Harmonic$density$ Registral$spread$ Inharmonicity$(harmonic$tension)$

Fig. 3.8. Parametric changes in the sixth section of Du cristal

movement is the foreground ornamentation based on rhythmic activity in the

percussion section.

At b.253 the marimba emerges from the arhythmic mesh of contrapuntal activity
with a repeated rhythmic motif built on quintuplet subdivisions. It is then joined by
the bass drum, timpani and roto-toms, each playing a different repetitive figure
based on triplet subdivisions. A simple repeated motif is played by the synthesizer,
piano, harp and vibraphone, based on the same tuplet grid as the bass drum. The
interlocking percussive material itself does not generate a great degree of tension.
The mid-tempo nature of the bass drums crochet tuplets acts as the dominant pulse
in this passage and is not fast enough to generate any sort of momentum-based
tension. This section relies on the aforementioned increase in dynamic profile for
this, with the rhythmic activity serving as foreground ornamentation while the two
main parametric changes, dynamic profile and harmonic density, converge to create
the sections tensional climax.

! 60!
SECTION 7 : bb.278329
The seventh section follows a familiar low-to-high tension trajectory, this time in
preparation for Du cristals final climax in the following section. Dynamic profile and
inharmonicity are again the main parameters affecting tensional fluctuations, sharing
a close relationship in this sections surges in tension (Fig. 3.9).

This section also starts with a change in harmonic field in its opening bars,
reinforcing the sense of transition to a new phase in the work. The seventh section
contains the most harmonic flux of the entire work, with two more changes to the
underlying harmonic field occurring. Though inharmonicity is initially relatively
high in this section (containing the highest peaks of inharmonicity in the entire work
between bb.278300), tension is initially minimized through the low dynamics that
follow the crescendo at the end of the previous section. Rhythmic activity appears
frequently, organized in a similar fashion to the short, arhythmic phrases described
in the first section. It is the periodic increases in dynamic profile and
inharmonicity/harmonic density that


278" 280" 282" 284" 286" 288" 290" 292" 294" 296" 298" 300" 302" 304" 306" 308" 310" 312" 314" 316" 318" 320" 322" 324" 326" 328"

Dynamic"pro5ile" Rhythmic"activity" Tensional"trajectory"

10$ !10$

0$ 0$
278$ 280$ 282$ 284$ 286$ 288$ 290$ 292$ 294$ 296$ 298$ 300$ 302$ 304$ 306$ 308$ 310$ 312$ 314$ 316$ 318$ 320$ 322$ 324$ 326$ 328$

Harmonic$density$ Registral$spread$ Inharmonicity$(harmonic$tension)$

Fig. 3.9. Parametric changes in the seventh section of Du cristal

! 61!
enhance the tensional qualities of the rhythmic activity. The dynamic surges take
place within a larger overall increase in dynamic profile that culminates at bb.299
300 with one of the loudest crescendi of the entire work a sustained statement of
the same chord that opens Du cristal, signaling a change in the underlying harmonic

From the opening bars of this section the use of glissandi in the string section
increases, until practically the entire string section is doing so by the time the large
crescendo arrives. The amount of pitch instability, and subsequent harmonic flux,
created by string glissandi contributes to the tensional trajectory in approaching the
large crescendo. The more synchronized relationship between surges in dynamic
profile and harmonic density/inharmonicity make this section somewhat unique,
and reminiscent of the close relationship between dynamic profile and registral
spread in the ninth section of Nympha, whose form is also based on a trajectory from
low to high tension in anticipation of climactic material. Following the sustained
pitches from the crescendo at bb.299300, rhythmic activity returns as foreground
ornamentation, but now with a foregrounded repetitive semiquaver grid in the
upper register of pitched percussion. Though harmonic density remains the same as
earlier in this section, harmonic tension is decreased. One further crescendo arrives
at b.315, though its tensional impact is lessened due to a decrease in harmonic
tension and an overall fall in dynamic profile following the main crescendo.
Emerging from this crescendo is more rhythmic activity, now transformed by a mesh
of short ascending phrases in the woodwind choir that grows increasingly
polymetric, also underlined by another change in harmonic field. The loss in
rhythmic stability is combined with another surge in dynamic profile to generate a
high degree of tension immediately before Du cristals last climactic passage in the
following section.

SECTION 8 : bb.330353
The penultimate section of Du cristal contains the final major climax of the work,
with dynamic profile and rhythmic activity dominating the tensional trajectory. The
overall form of this section follows a transition from high to low tension, as both
rhythmic activity and dynamic profile subside, creating a rest period ahead of the
works final section (Fig. 3.10).

The crescendo at the end of the previous section bursts into a rhythmic passage that
marks the beginning of the eighth section. The dynamic profile leaps to ff and the
familiar insistent semiquaver rhythm returns, though in this instance the rhythmic

! 62!

330" 331" 332" 333" 334" 335" 336" 337" 338" 339" 340" 341" 342" 343" 344" 345" 346" 347" 348" 349" 350" 351" 352" 353"

Dynamic"pro6ile" Rhythmic"activity" Tensional"trajectory"

10$ !10$

0$ 0$
330$ 331$ 332$ 333$ 334$ 335$ 336$ 337$ 338$ 339$ 340$ 341$ 342$ 343$ 344$ 345$ 346$ 347$ 348$ 349$ 350$ 351$ 352$ 353$

Harmonic$density$ Registral$spread$ Inharmonicity$(harmonic$tension)$

Fig. 3.10. Parametric changes in the eighth section of Du cristal

material is quickly passed around the percussion section, most notably between the
xylophone, doubled by marimba and piano, and roto-toms, doubled by timpani. As
found in a handful of passages in Du cristal, this passage sees the percussion section
and the rest of the orchestra share the same metric grid, momentarily bringing the
work into sharp focus. This passage is the most dynamic and foregrounded iteration
of the orchestras metric synchronisation, as the two groups of doubled percussion
instruments quickly trade small groupings of semiquaver patterns to create an
almost unbroken stream of semiquaver beats. Rhythmic activity is also reinforced in
this passage through bursts of brief sustained notes from other instruments,
particularly the brass section. The overall effect is one of high rhythmic tension, as
the simple underlying metre is subverted due to the unpredictable passing of
rhythmic fragments between the two groups of doubled percussion instruments and
the additional unpredictability of the brass and woodwind bursts.

While harmonic density is only slightly less than that of the previous section,
inharmonicity falls to some of its lowest levels of the entire work. This is another
example of Saariaho subverting parametric relationships, once again not relying on
inharmonicity to generate tension. The transition to the eighth section is again

! 63!
marked by a change in the underlying harmonic field. It is apparent that changes in
the underlying harmonic field in Du cristal, which, in the atonal context of this work
can be difficult to perceive, are supported by parallel changes in other parameters to
give these harmonic transitions structural significance.

Once the rhythmic material fades away, the remainder of the eighth section serves as
a low-tension rest area dominated by the extensively divided string section. This
passage shares the subversive combination of parametric extremes noted in the third
sections string-dominated rest area, with typically climactic parametric features
such as wide registral spread and high harmonic density juxtaposed with quiet
dynamics. Between its registral extremes (the widest of the entire work along with
the sixth section), however, this passage also contains more internal movement than
its earlier counterpart, with glissandi in many string parts generating a muted degree
of harmonic flux and a multitude of short phrases creating another dense polymetric
mesh. Nevertheless, the low dynamics restrain the material, providing a platform for
the segue into the quiet suspense of Du cristals final section.

SECTION 9 : bb.354368
The final section of Du cristal is unique for its tensional process, relying on repetition
as a means of sustaining tension as the work winds down, with only the string and
percussion sections remaining. Synchronisation is also turned into a foreground
feature, becoming a final structural point of arrival for the section and the work as a
whole (Fig. 3.11).

A slight increase in harmonic tension occurs at the beginning of the section but this
increase is not sufficient to stand out as strongly inharmonic in the overall context of
Du cristal. Furthermore, inharmonicity immediately decreases throughout the
section. Also gradually decreasing is dynamic profile, adding to what would usually
be a release of tension. Tension remains, however, due to the percussive material
occupying this sections foreground ornamentation a slow pulse played on the
bass drum, with each beat quickly echoed by a high C on the crotales. The extreme
registral spread of the string section remains unchanged from the previous section,
contributing to the climactic quality of the harmonic material. The feeling of
suspension is enhanced through the absence of glissandi in this section, leaving
pitches hanging, with trills being the only form of pitch movement. Although a
reliable periodicity exists in the bass drums rubato pulse, it is a descriptive stretch to
regard it as rhythmic activity. Still, the periodicity of the bass drum/crotales pulse
must be acknowledged as a tensional device, its repetition sitting alongside the

! 64!

354" 355" 356" 357" 358" 359" 360" 361" 362" 363" 364" 365" 366" 367" 368"

Dynamic"pro6ile" Rhythmic"activity" Tensional"trajectory"

10$ !10$

0$ 0$
354$ 355$ 356$ 357$ 358$ 359$ 360$ 361$ 362$ 363$ 364$ 365$ 366$ 367$ 368$

Harmonic$density$ Registral$spread$ Inharmonicity$(harmonic$tension)$

Fig. 3.11. Parametric changes in the ninth section of Du cristal

shimmering strings as the final section exists suspended in time. It is rhythmic

synchronization in its most minimal form that finally brings closure to Du cristal.
After the desynchronized bass drum/crotales pulse is repeated for almost a minute,
the final two pulses see the two instruments synchronise (b.364), falling quietly on
the same beat.

To summarise, given that the sectionalisation of the works in this analysis is based on
the identification of small to mid-scale tensional trajectories that take place within
their overall durations, the relatively small number of discrete sections identified in
the analysis of Du cristal reveals the slow rate at which tensional changes tend to take
place in the work. That the transitions to new sections in Du cristal often coincide
with a transition to a new underlying harmonic field suggests that harmonic change
is not employed in varying speeds to generate tensional changes in Du cristal, but
provide harmonic variety. This approach to harmonic change differs from Nympha
in that while it is common for a new section to start with a new harmonic field, the
rate of harmonic change is also used within a section as a device for reinforcing
short-term tensional changes.

! 65!
As with Nympha, parametric desynchronisation dominates Du cristal. Saariaho
avoids combining the tensional extremes of many parameters at once, choosing to
subvert the climactic potential of a passage. Structurally, both Nympha and Du cristal
use rhythmic synchronisation to underscore climactic points of arrival.

Changes in the sound/noise parameter do not greatly affect the overall tensional
state in the orchestral context of Du cristal, where techniques such as extreme bow
pressure are now confined to just one section of a larger ensemble. Saariaho does not
attempt to augment the presence of noise in Du cristal by incorporating extended
techniques into other instrument parts, perhaps because incorporating noise-based
techniques into other instrument parts would compromise the conceptual realization
of the work. Speculation aside, the absence of noise as a tensional device is the
primary parametric point of difference from Nympha.

Without the sound/noise continuum to rely on for creating tensional change,

inharmonicity assumes the role of the parameter most closely tied to changes in
tensional state, alongside dynamic profile. The sheer density of Du cristals harmonic
material results in far higher levels of inharmonicity than Nympha. It is difficult,
however, to quantify and compare the two works and identify one work that is more
inharmonic than the other, as both works possess only extremely small amounts of
harmonic material that reach beyond a zero value for inharmonicity, residing instead
in a constant state of inharmonicity. As discussed, wide registral spread surely plays
a role in distorting the perception of harmonic tension in Du cristal perhaps more
so than in Nympha due to some intervals spanning multi-octave distances.

On top of the dual role of rhythmic activity in Du cristal as 1) a structural point of

arrival (tutti) and 2) as provider of contrapuntal detail that contrasts with the works
globalized textures, there are moments in Du cristal that see rhythmic activity cast as
an overtly tensional device while serving as foreground ornamentation, such as the
climactic timpani/roto-tom solo passage in the fifth section (where inharmonicity is,
incidentally, rather low).

! !

! 66!
for alto flute, cello and piano

Composed eight years after Du cristal and 11 years after Nympha, Cendres (1998)
displays a detectable stylistic departure from the other two analysed works. Cendres
is, however, still able to be divided into sections based on tensional states or
transitions between two or more tensional states (Fig. 4.1). Cendres is distinguishable
from the other two analysed works due to the differing use of several parameters,
while its shorter duration necessarily allows for less of the gradual parametric
change employed in Nympha and Du cristal.

Given the shorter duration of Cendres and the faster tensional transitions that take
place within the work, the need for foreground ornamentation material that
provides moment-to-moment interest while larger formal and tensional processes
unfold is somewhat reduced. As with Du cristal, noise plays little part in
controlling tensional states, occasionally serving as foreground ornamentation but
not in the overpowering manner observed in Nympha. Harmonic density is also
arguably used as foreground ornamentation in passages involving solo instruments.

Of the three analysed works, the use of rate of harmonic change as a large-scale
tensional device is most audible in Cendres (Fig. 4.2, 4.3). While the rate of harmonic
change increased in Nymphas climactic passages, the overpowering noise produced
by extreme bow pressure and other bowing techniques often obscured pitch content.
In Cendres, the relatively pure quality of its pitch content allows changes in the
underlying harmonic field to be more easily perceived, though pitch instability often
keeps inharmonicity levels in a constant state of flux. The use of rhythmic unison as a
structural point of arrival is present in Cendres, a parametric role that stylistically
links all three works.
SECTION 1 : bb.124
The opening section of Cendres follows a clear trajectory from low to high tension,
with the parameters that contribute most to this tensional change being dynamic
profile and inharmonicity. While the minimal harmonic density of Cendres opening
bars necessarily results in the foregrounding of one instrument at a time, once the
entire trio is playing no one instrument provides foreground ornamentation. This is
similar to the globalised textural material found in much of Nympha and Du cristal.
Though material that could be considered foreground ornamentation is elusive in!

! 67!
Section Bars Time Tensional Main features

1 124 0:001:43 Low to high Low harmonic density initially. Increases

in dynamic profile, registral spread and

2 2566 1:444:11 High to low Pitch instability techniques, introduction

of rhythmic activity via polymetric
passages, gradual reduction of tension
via transition to more consonant
harmonic material.

3 6794 4:125:08 Low to high, Rhythmic unison at tensional high-

with points, frequent alternation between two
fluctuations in parametrically distinct types of material,
between especially in terms o harmonic tension.

4 95136 5:097:08 Low/mid Rhythmic activity via polymetric

tension material, predominantly low harmonic
density, smaller registral spread and
lower dynamic profile.

5 137154 7:098:20 Low to mid, Repeated figure appears in flute and

with three cello. Low harmonic density, accelerating
distinct rate of harmonic change.
tensional surges

6 155168 8:219:00 High tension Wide parametric and tensional

fluctuations, highly accelerated rate of
harmonic change.

Fig. 4.1. Formal outline of Cendres

this section, when it does appear it is often through changes in the sound/noise
parameter, as each instrument employs some form of extended technique that alters
timbral quality. Additionally, harmonic density can be said to assume the role of
foreground ornamentation in the opening bars of the work, as the textural sparseness
of the material draws attention. There is also little foregrounded rhythmic activity in
the first section. As with Nympha and Du cristal, rhythmic activity that indicates an
underlying pulse is used sparingly but often contributes to the illumination of formal
landmarks in the work.

! 68!


0.0" 1:00" 2:00" 3:00" 4:00" 5:00" 6:00" 7:00" 8:00" 9:00"



0:00" 1:00" 2:00" 3:00" 4:00" 5:00" 6:00" 7:00" 8:00" 9:00"


Fig. 4.2. Rate of harmonic change and tensional trajectory in Cendres

Fig. 4.3. Sequence of harmonic fields in Cendres, mapped against clock time and sectional

! 69!
This first section of Cendres is notable for a relatively high number of parameters
changing in parallel as the tensional transition takes place (Fig. 4.4). The work opens
with a single pitch from the piano (minimal harmonic density and registral spread),
played mp (low dynamic profile). Meanwhile, there is no change in the underlying
harmonic field until b.12, by which time these parameters have increased to produce
a slightly stronger degree of tension. It is the change in harmonic material that
significantly transforms the amount of tension in this section, introducing strongly
inharmonic material for the first time in the work. Previously, harmonic material has
been mostly consonant. Fast and resonant piano runs, containing chromatic passing
notes that are not part of the underlying harmonic field, now contribute to the sense
of denser harmonic material being used.


1" 2" 3" 4" 5" 6" 7" 8" 9" 10" 11" 12" 13" 14" 15" 16" 17" 18" 19" 20" 21" 22" 23" 24"

Dynamic"pro6ile" Harmonic"density" Registral"spread" Tensional"trajectory"

Fig. 4.4. Parametric changes in the first section of Cendres

The more synchronised parametric change in this section is demonstrated a few bars
later, between bb.1517, when a brief period of release occurs. Dynamic profile
decreases, registral spread significantly decreases to less than an octave and
inharmonicity decreases. Foregrounded rhythmic activity appears for the first time at
this point also, as the players pass a repeated three-note demisemiquaver pattern to
one another in bb.1617. As mentioned, this period of release is a brief interruption
of the overall trajectory of this section from low to high tension dynamic profile
and inharmonicity return to their former levels and registral spread in particular
expands to its widest range yet, adding to the sense of climactic tension at the end of
the section. Pitch instability also features strongly in the final bars of the section
through the repeated use of glissandi in the flute part alongside trilling in the cello
and piano parts.

The parallel parametric changes also roughly coincide with the first change in the
underlying harmonic field. The rate of harmonic change increases as a new

! 70!
underlying field emerges with the release of tension at bb.1516, with another
change in harmonic material used a few bars later with the transition back to the
high-tension conclusion of the section.

Although a degree of noise is present in the opening material of this section, it is not
sufficient to act as a source of tension, functioning more as foreground

ornamentation. The Eb2 that opens Cendres, played on piano and then cello,
incorporates extended techniques on both instruments, introducing foreground

ornamentation via timbral manipulation. The pianos Eb2 in the opening bar is
played by plucking the string inside the piano, producing an attack profile
resembling a pizzicato articulation on a string instrument. The following Eb2 from
the cello is played sul pont (gradually transitioning to naturale), while trilling with
harmonic finger pressure to produce G4 (the fifth partial of the harmonic series based
on Eb2). The breath tones played on the flute serve as foreground ornamentation
when it enters at b.17.

SECTION 2 : bb.2566
The second section follows an opposing trajectory to the first section, moving from
high to low tension. As with the first section, parameters often share parallel
relationships with one another, with more contrasting material in terms of high and
low tension occurring during the overall tensional process of the section (Fig. 4.5).
Dynamic profile and inharmonicity are again the primary parameters used to
generate tension, with the pitch instability of techniques such as trilling and
glissando also continuing to affect tension.


25# 26# 27# 28# 29# 30# 31# 32# 33# 34# 35# 36# 37# 38# 39# 40# 41# 42# 43# 44# 45# 46# 47# 48# 49# 50# 51# 52# 53# 54# 55# 56# 57# 58# 59# 60# 61# 62# 63# 64# 65# 66#

Dynamic#pro6ile# Harmonic#density# Registral#spread# Rhythmic#activity# Tensional#trajectory#

Fig. 4.5 Parametric changes in the second section of Cendres

! 71!
The second section begins with the most dominating use of noise thus far in Cendres,
as the flautist uses loud breath tones, mouthing a series of phonemes through the
instrument. The use of excited human breath certainly adds dramatic tension to the
beginning of the section the sound of human breath, not being controlled through
embouchure and transformed into pure pitch, instils a more instinctive sense of
tension for the listener, not unlike the effect of an increase in dynamic profile as
discussed in the methodology section. Whereas pitch-based tension and release is to
some extent a learned interpretation of sound, animated unpitched vocal material
and high amplitude of a sound call on the immediacy of our biological reaction to
extramusical environmental phenomena.

Harmonic tension varies widely between inharmonicity and more consonance

throughout most of this section, but pitch instability is the more relevant tensional
device employed. The large amount of trilling and glissandi that occupy the
instrument parts in this section makes reliable calculation of harmonic tension
difficult. At almost any point in this section at least one of the instruments is
performing a trill, a glissando or a fast ascending run comprised of pitches outside of
the underlying harmonic field, causing much of the section to be dominated by
harmonic flux.

Although the general trajectory of the section is from high to low tension, this is
interrupted on several occasions by the freezing of certain parameters in combination
with repetition of material by the whole trio, breaking away from the arhythmic,
harmonically unstable material that dominates the section. In bb.4142 each
instrument switches to a repeated figure, temporarily stabilising registral spread,
dynamic profile and harmonic density. The most noticeable parametric change,
however, is the sudden prominence of rhythmic activity created by the polymetric
nature of the repeated material. This is the first instance across the three analysed
works that the use of rhythmic activity as foreground ornamentation is manifested in
this style. When the ensemble unites using rhythmic activity in Nympha and Du
cristal, all instruments repeat the same rhythmic material, with little or no pitch
variation in the individual lines.

The repeated polymetric material that follows from bb.5056 (Ex. 2.1) provides even
more contrast to the surrounding material due to its relative lack of harmonic
tension. The strongest source of tension in the two passages of rhythmic activity is
the push and pull of the polymetric material as the distance between the respective
downbeats of the flute and cello patterns incrementally change. Like the material in

! 72!
Ex. 2.1. Polymetric material in the second section of Cendres!!

bb.4142, this short passage sees the parameters freeze in another structural non
sequitur before the work reverts back to its stylistic character, as though these
tensional departures might never have happened.

Despite the numerous ascending runs superimposed over the pitch material, this
section sees little actual change to the underlying harmonic field. The section retains
the harmonic field from the end of the first section, transforming amongst the
extended series of piano runs from bb.3540 and emerging with a new underlying
field for the first of the short repetitive passages at b.41. Following the second
repetitive passage, the transition to a low-tension state is completed through
harmonic density gradually thinning out, accompanied by a decrease in dynamics.

The harmonic field is ultimately reduced to the consonance of Eb2 and G4, revealing
that this has been a return to the harmonic field that opens Cendres.

! 73!
SECTION 3 : bb.6794
While following an overall trajectory from low to high tension, the brief third section
contains several large tensional fluctuations throughout. Rhythmic activity is
harnessed in the same manner seen in Nympha and Du cristal, with the ensemble
synchronising to create rhythmic unison in passages of high tension. Dynamic profile
is again closely tied to tension. The behaviour of the inharmonicity parameter is of
interest in this section due to its unusual relationship with harmonic density.

The transition between the second and third section is marked by the full bar rest at
b.66 and a new underlying harmonic field starting at b.67. The same harmonic field
remains in place for the entire section. A significant amount of tension is generated
in this section through the frequent alternation between two parametrically distinct
types of material. The first type of material is typical of what has preceded this
section arhythmic, with frequent odd-numbered beat subdivisions, set against
sustained and trilled notes and is associated more with moments of lower tension
in this section. The second type of material found in this section uses rhythmic
unison in the form of repeated semiquaver patterns, with minimal harmonic density
and registral spread (Ex. 2.2). The material containing rhythmic unison also contains
the most inharmonic material, and is used for moments of higher tension in the
section. This section contains the most foregrounded rhythmic activity and rhythmic
unison thus far in the work, signalling a formal point of arrival.

The semiquaver pattern is more pervasive nearer the end of the section, providing a
sense of tensional momentum. The structure of the section is reinforced by the
cadential use of the C2 in the final chord, acting as a point of arrival for the section.
There is little tension derived from inharmonic material in the third section, with the
majority of harmonic material actually measuring as consonant (Fig. 4.6). The

Ex. 2.2. Rhythmic unison in the third section of Cendres!

! 74!

67" 68" 69" 70" 71" 72" 73" 74" 75" 76" 77" 78" 79" 80" 81" 82" 83" 84" 85" 86" 87" 88" 89" 90" 91" 92" 93" 94"

Dynamic"pro6ile" Registral"spread" Rhythmic"activity" Tensional"trajectory"



67# 68# 69# 70# 71# 72# 73# 74# 75# 76# 77# 78# 79# 80# 81# 82# 83# 84# 85# 86# 87# 88# 89# 90# 91# 92# 93# 94#





Fig. 4.6. Parametric changes in the third section of Cendres

relationship between harmonic density and inharmonicity in this section is of interest

due to this lack of harmonic tension even in the sections more dense harmonic
material a contrast with the usual relationship between these two parameters in
Nympha and Du cristal. Three of Cendress largest spikes in harmonic density occur
in this section, in the form of three thick piano chords at b.73, b.79 and bb.8283.
Each of these chords, however, measure as consonant. This consonance compliments
the mostly sustained note durations occurring with these chords, supporting
moments of small-scale tensional release in the section. Contrasting with this is the
inharmonicty of the high-tension rhythmic material that emphasises the small-scale
movement between higher and lower states of tension during the section. The overall
transition from low to high tension in this section is reinforced by the prevalence of
the inharmonicity associated with the high-tension material in the final few bars.

One explanation for the relative consonance of this sections harmonic material could
be the absence of quarter-tones in much of Cendres, meaning that the occurrence of
highly dissonant intervallic values between equal-tempered and quarter-tone pitches
is reduced. Pitch instability also features in this section, predominantly in the form of
trilling, which again makes the calculation of inharmonicity difficult in some bars.

! 75!
SECTION 4 : bb.95136
Following the frequent tensional fluctuations and high-tension conclusion of the
third section, the fourth section displays more tensional stability, particularly
through occupying a smaller dynamic range. Although most of its material can be
considered low to mid tension, there is enough tensional and parametric variation,
including two short dynamic surges, to keep this section from being considered a
rest area of the kind described in analyses of Nympha and Du cristal (Fig. 4.7).
Tensional changes are again driven by movement in the dynamic, inharmonicity and
rhythmic activity parameters. The change to a new harmonic field around b.108
marks the beginning of a slightly accelerated rate of harmonic change in Cendres,
though in this section the acceleration is not sufficient to create tension through any
sense of harmonic flux.

The fourth section retains the predominantly low harmonic density of the third
section, evading any sense of vertical mass that might suggest climactic material. In
fact, it is rare in this section for the whole trio to be playing at the same time, with the
only clear increases in density occurring in bb.115118, where the pianos chordal
material thickens or the cello plays double stops, and b.125, which reprises the thick
piano chord found in the third section. The amount of inharmonicity is also similar
to the third section, although this section is more consistently inharmonic from its
beginning at b.95 until b.108, where the change to less consistent inharmonicity
roughly coincides with a change to a new underlying harmonic field. As also
observed in the third section, harmonic density does not share a parallel relationship
with inharmonicity. Both of this sections peaks in harmonic density bb.115118
and b.125 coincide with almost entirely with consonant pitch collections,
subverting the typical tensional relationship between these two parameters. Pitch
instability is also employed regularly in this section, appearing again in the form of
trilling and glissandi. The pitch instability generated by these techniques continues
to play an important role in the tensional quality of Cendres by producing an almost
constant state of restlessness within the harmonic fabric of the work.

As with harmonic density, the registral spread of the fourth section often restricts the
cadential and tensional range. Of the six sections that make up Cendres, the fourth
section contains the least amount of fluctuations of registral spread, being
predominantly confined to the middle and upper registers. The absence of pitch
material in the lower bass register is the most noticeable feature regarding registral
spread in this section, contributing to its structural identity as a predominantly
lower-tension area in the overall context of the work, without a strong climactic

! 76!
event. The bass register is used in the opening four bars, producing the widest
registral spread of the section (four octaves), but Saariaho subverts its climactic
potential by avoiding rhythmic activity, limiting harmonic density and using a low
dynamic profile. According to calculations, some of the harmonic material in the
opening four bars is consonant. This consonance is neutralised to a significant degree
by the wide registral distribution of the constituent pitches, producing the more
consistent harmonic tension referred to in the previous paragraph.


95# 96# 97# 98# 99# 100# 101#102#103#104#105#106#107#108# 109#110# 111#112#113#114#115#116#117#118# 119#120# 121#122#123#124#125#126#127#128# 129#130# 131#132#133#134#135#136#

Dynamic#pro6ile# Registral#spread# Rhythmic#activity# Tensional#trajectory#

Fig. 4.7. Parametric changes in the fourth section of Cendres

In this section, rises in tension are achieved by combining increased rhythmic activity
with high dynamic profile, along with inharmonicity that is generated to some
degree through the expansion of registral spread. The use of rhythmic activity as a
tensional device is somewhat unique in this section, as rhythmic unison between the
players is not used, though an underlying metric grid is still revealed. The first rise in
tension, between bb.11518, sees the cello as the sole source of rhythmic momentum,
playing a quintuplet semiquaver pulse over tremolos in the piano part. The second
rise in tension, in bb.126129, sees both the flute and piano playing consistent
semiquavers, though after the four semiquavers on a unison C#4 on the first beat of
b.126 (the only example of rhythmic and pitch unison in this section), rhythmic
unison is made evasive through the use of opposing beat subdivisions in the flute
part while the cello holds the initial semiquaver subdivision. In both of these
examples Saariaho creates momentum, and subsequently tension, through the use of
a repetitive pulse, though tension is controlled or curbed through not having
rhythmic activity dominate the material through complete rhythmic unison. The
result is that the climactic impact of the material is diminished and is not
convincingly perceivable as a clear point of arrival in the formal structure of the

! 77!
Again, no one parametric feature is foregrounded in a fashion that could be
considered foreground ornamentation. Rather, the short, expressive phrases that
emerge from this sections sparse textures provide a relatively conventional
foreground role, such as the cello line in bb.108109 and bb.121122, and the flute
line in bb.119120. The repetition of phrasal material, such as the aforementioned
cello lines, serves more as a structural, formal device, providing coherence for the
section as opposed to providing small-scale detail while larger parametric
trajectories unfold (as observed in Nympha and Du cristal).

SECTION 5 : bb.136154
As with the fourth section, the fifth section contains more controlled tensional
fluctuations, this time in preparation for the climactic sixth and final section of
Cendres. The section is comprised of three small-scale trajectories from low to mid-
tension, with the third small-scale tensional trajectory breaking into the high-tension
material that opens the sixth section (Fig. 4.8). Dynamic profile and inharmonicity
are again the drivers of tensional change in the section. Rhythmic activity is
foregrounded in low-tension passages in this section, in contrast to its typical use in
the three analysed works as an additional, structurally significant device for high-
tension passages.

Beginning each low-to-mid tension trajectory in the fifth section is a bar of short
ascending and descending glissandi in the flute and cello parts. Found in bars 136,
147 and 152, the flute and cello material is identical in all three bars (Cendres is the
only work of the three containing identical repetition of motivic material). The flutes
unbroken line of six quavers at =50 provides a rhythmic steadiness that lacks the
momentum of the foregrounded rhythmic material in Cendres that is typically based
on much faster semiquaver subdivisions. In combination with the mf/mp dynamic
markings, this material provides a low-tension foundation from which tension
increases in subsequent bars before ebbing back to the same low-tension motif.

The rate of harmonic change in this section is at its fastest thus far in the work, as the
mid-tension material between bb.138146 contains changes to two different
harmonic fields before reverting back to the harmonic field carried over from the end
of the fourth section for the second statement of the low-tension flute and cello
material. Though the rate of harmonic change increases in this section, it is really the
cyclic pacing of the repeated small-scale tensional trajectories that generates the
sense of accelerating forward movement the first small-scale tensional trajectory
occupies bb.136146, while the second occupies bb.147151 and third b.152 until

! 78!

137" 138" 139" 140" 141" 142" 143" 144" 145" 146" 147" 148" 149" 150" 151" 152" 153" 154"
Dynamic"pro6ile" Harmonic"density" Rhythmic"activity" Tensional"trajectory"
Fig. 4.8. Parametric changes in the fifth section of Cendres

b.155, when the high-tension final section erupts and harmonic flux truly becomes a
strong tensional device. The underlying rate of harmonic change in the fourth section
is actually rather constant beneath the increasingly shortened small-scale trajectories
involving other parameters such as dynamic profile and inharmonicity.

As in other low-tension sections in Cendres, harmonic density remains fairly low

throughout this section, increasing by one or two pitches whenever the piano
extends its range below the bass clef, greatly extending registral spread. Low
harmonic density again supports the low-to-mid tensional state, and also diminishes
the climactic impact of this sections expansions of registral spread, maintaining a
wide spacing between pitches reminiscent of the muted climactic material in

SECTION 6 : bb.155168
The brief final section of Cendres is the tensional high-point of the work, achieved as
much through the quick changes between states of high and low tension as through
the high-tension material itself. The accelerating rate of harmonic change reaches its
peak in this section, contributing to the contrast between high and low tension
material. The dynamic profile and inharmonicity parameters also fluctuate greatly in
this section, maximising the sense of flux.

The section opens with the full trio interrupting the pianos gradual dynamic surge
from the end of the previous section, drastically increasing the tensional state with
an arhythmic flurry of pitches from a new harmonic field, played forte with a mixture
of trilling and glissandi in the flute and cello and a brisk phrase comprised of
demisemiquavers in the piano part. The following bar sees an immediate tensional
change, achieved through several simultaneous parametric changes in a relatively

! 79!
conventional manner. Harmonic density reduces down to single pitches from the
flute and cello, their dynamic markings dropping to mf and mp respectively. Perhaps
the most striking parametric change is harmonic, with the pitch content reprising the

sustained major third on Eb from the works opening harmonic field. Also
contributing to the sense of flux here is the sudden reduction in registral spread

from the three and a half octaves in b.155, the Eb major dyad in b.156 is spread across
a major tenth in the bass clef.

Similarly contrasting material follows for the remainder of the section. Aside from
b.164, however, the full trio no longer plays at the same time, instead alternating
between solo piano and flute and cello, with the high-tension material simply a
repeat of the pianos short phrase from b.155 (Ex. 2.3). The flute and cello, dividing

Ex. 2.3. Repeated material in the sixth section of Cendres

! 80!
the pianos hands between them, use glissandi to further increase tension through
pitch instability. This instrumental variation is another device for maximising flux,
and subsequently tension, via constant timbral alternation. A side effect of this
sections instrumental alternation is that harmonic density remains very low, in
another apparent parametric subversion of climactic material (Fig. 4.8). The
overpowering tensional flux is, however, such that the low harmonic density is
somewhat negated. This negation is also achieved through the predominantly wide
registral spread (another powerful climactic device) in this section, realised by the
wide intervallic leaps used in each part.

The short phrase repeated throughout this section is deceptively rhythmic this is
made clear when it appears in the solo piano part at bb.161163 (the only time the
phrase appears in the piano part unaccompanied), with the sharper, more percussive
attack and timbral homogeneity of the piano providing clearer rhythmic focus than
the flute/cello iterations peppered with glissandi. Here, rhythmic activity is briefly
brought to the foreground in a unique incarnation: there is metric consistency in the
demisemiquaver subdivision, which is free of any of the metrically subversive tuplet,


155" 156" 157" 158" 159" 160" 161" 162" 163" 164" 165" 166" 167" 168"
Dynamic"pro6ile" Harmonic"density" Rhythmic"activity" Tensional"trajectory"


155# 156# 157# 158# 159# 160# 161# 162# 163# 164# 165# 166# 167# 168#




Fig 4.8. Parametric changes in the sixth section of Cendres

! 81!
quintuplet or septuplet subdivisions that have typically accompanied such material
elsewhere in the work. Here, in this brief snatch of time, is a slight loosening of
Saariahos grip on rhythmic behaviour in Cendres, with sense of freer rhythmic play,
without venturing into syncopation. Its brevity precludes it from becoming a
climactic point of arrival of great structural significance it seems, instead, to be
just one of several small-scale ingredients that contribute to contrasting character of
this section in the context of its overall scheme.

Continuing with reflection on the formal, structural significance of this section in the
overall context of Cendres, we should return to the aforementioned observation that
this short section is essentially comprised of alternations between two contrasting
types of tensional musical material, which presents in itself an obvious contrast to
the rest of the work. The clearest impression of a long-term parametric trajectory in
Cendres is found when reviewing the consistent acceleration of the rate of harmonic
change across the duration of the work, which reaches its apex with the frequent
change of harmonic field in the final 14 bars of the work. In this light, the sixth
section appears to function as the parametric climax for the rate of harmonic change,
while appearing in other regards as a coda or structural non-sequiter of the kind
referred to in the discussion of Nympha.

! !

! 82!


I will now discuss the findings from the analyses of Nympha, Du cristal and Cendres,
with a highlighting of the main similarities and differences in Saariahos use of
various parameters, and combinations of parameters, to change and maintain
tensional states across the three works. A description of how perceivable form is
created through these parametric changes is also given, along with a discussion of
the similarities, differences and functions of foreground ornamentation across the
three works.

Starting with a broad perspective of these works, I will discuss musical form as it
relates here. The perception of form for these works clearly does not occur through
the traditional analytical method of identifying sections that contain literal repetition
of a previous section, perceivable through repeated chord progressions or changes in
key. While changes in the underlying harmonic field commonly coincide with the
beginning of a new section, it is, first and foremost, the cycles of tension and release
that define the formal character of these works. It should be noted that all three
works contain at least one instance of chordal material being repeated at a later point
in the work. While they provide rare landmarks of coherence and familiarity, these
literal repetitions are too fleeting to define the formal character of the works in any
way. The sections that comprise the overall form of each work whether these
sections navigate trajectories from one tensional state to another or maintain a
constant tensional state throughout are distinguishable by their unique
combinations of parametric properties, providing variation across the duration of
each work and reinforcing an overall formal structure.

By changing which parameter, or combination of parameters, drives tensional

change, Saariaho retains variety as the works unfold, allowing for sections that
contain similar tensional states but through different parametric combinations. This
technique is behind the muted climactic material that features, in varying quantities,
in each work, whereby one parameter seems out of step with the tensional trajectory
of its surrounding parameters by being held at its low-tension extremity, subverting
what would otherwise be a more high-tension passage (as observed in the eighth
sections of Nympha and Du cristal and the fifth section of Cendres).

! 83!
The first parameter to be discussed is rhythmic activity. There is a clear relationship
between rhythmic activity, in the form of rhythmic unison, and states of high tension
in all three works (making rhythmic activity most associated with states of high
tension). Saariaho reliably uses rhythmic unison as a point of arrival, heightening
climactic material through the forward momentum generated by the whole ensemble
playing repeated pitches, typically of very short duration such as semiquavers. Much
of the impact of rhythmic unison in these works is due to its strong contrast with the
material that typically surrounds it rhythmic activity, in which an underlying
pulse is perceivable, is absent for the vast majority of each work. An underlying
metric grid is revealed only occasionally on these occasions, the temporal
dimension of the music is engaged in a more active (rhythmic) fashion, giving the
impression of acceleration to the processes unfolding in the works and heightening
the sense of tension and momentum.

Of all the parameters traced through the selected works, rhythmic activity is also the
most versatile. It is employed not only in the form of rhythmic unison for passages of
heightened tension, but as more playful foreground ornamentation that provides
moment-to-moment interest while underlying tensional trajectories unfold.

Occasionally, freer rhythmic material is employed as a source of tension, as observed

with the soloistic percussion material in Du cristals climactic fifth section. Here, the
percussive material functions as both foreground ornamentation and the primary
source of tension in the section, due to its high amplitude. Along with sound/noise,
rhythmic activity is the main parameter used for foreground ornamentation in
Nympha, and it is the primary parameter used for foreground ornamentation in Du
cristal. A freer type of rhythmic activity is also applied in Nympha when it appears as
the proto-melodic material that acts as foreground ornamentation, though this
typically appears in non-climactic passages.

Cendres differs from Nympha and Du cristal because of its use of looping
polyrhythmic and polymetric material that lasts for several bars, for example in its
second section. These occasional snatches of rhythmic material do not develop to
become an important feature in the work, nor are they part of climactic material,
instead serving as non sequiturs of sorts, brief moments of rhythmic contrast.

The rate of harmonic change is another defining feature of the selected works. In
considering the relationship between the rate of harmonic change and the perception
of form in these works, the coincidence of changes in the underlying harmonic field

! 84!
with a transition to a new section of tension and release shows that harmonic change
is important to Saariaho for reinforcing a sense of structural development and
forward movement on a sectional level, and less as a moment-to-moment tensional
device via accelerating its rate of change. While rate of harmonic change appears to
be a parameter that is used to support a handful of short-term tensional changes in
Nympha and Du cristal, it follows a long-term trajectory for the duration of Cendres,
moving from an initially slow rate of harmonic change to the frequent shifts between
harmonic fields in its final section.

Across the three works, however, rate of harmonic change is very rarely fast enough
to create a genuine sense of flux and serve as a source of high tension. The only
sections in which Saariaho uses an accelerated rate of harmonic change to build
tension are the sixth section of Nympha, which contains the first major climax of the
work, and the final, climactic section of Cendres. As mentioned in the analysis of the
sixth section of Nympha, however, the noise generated by extreme bow pressure
obscures the increased rate of harmonic change taking place. It is more common for
Saariaho to use a slow rate of harmonic change to support periods of low tension or
stasis in these works, such as in third section of Du cristal.

Inharmonicity is also a key parameter in the production of tension across all three
works. While Du cristals harmonic material is measured as entirely inharmonic,
Nympha and Cendres contain brief flickers of more consonant material, though it is
only in Cendres that we find what may be termed intentional consonance, in the
harmonic material based on the Eb2 that opens the work. Registral spread plays an
important role in the perception of inharmonicity. As mentioned in the analyses of
the selected works, wide spacing between a chords constituent pitches can
compromise the perception of its intervallic relationships, resulting in the
consonance or dissonance of harmonic material becoming less audible. This effect
certainly comes into play in each of these works, arguably a device employed by
Saariaho to give the impression of fluctuation in the degree of inharmonicity,
creating a type of harmonic ambiguity. While inharmonicity is an important
tensional parameter here, the most climactic, high-tension sections in these works are
not always the most inharmonic. It is, unsurprisingly, the dynamic profile parameter
that truly defines climactic material in these works. Du cristal is by far the most
inharmonic of the three works, based on the measurements of inharmonicity. This is
largely due to passages of high harmonic density, often containing dissonant
quarter-tone intervals. The climactic fifth section, however, with its thunderous
percussive material, contains some of the least inharmonic material of the entire

! 85!
work, mostly due to it containing far less quarter-tones. It seems that since these
works constantly dwell in inharmonicity, Saariaho must make use the tensional
extremes of other parameters such as rhythmic activity and dynamic profile to
realise tensional peaks.

Also tied to the aforementioned prevalence of inharmonicity in the three works is

Saariahos use of pitch instability to generate tension, which leads to instability of the
intervallic relationships within harmonic material in the context of ensemble works.
This is achieved through extensive use of trills, glissandi and, in the case of strings,
constantly shifting bow positions. This method of producing tension through pitch
and harmonic instability is evident in the prevalence of these playing techniques in
Nympha. Glissandi are used in the flute and cello parts throughout Cendres, while all
three instruments regularly employ trills. The cellos changing bow positions in
Cendres are reminiscent of Saariahos string writing in Nympha. Du cristal also
features this type of pitch instability in the string section, while trills are used
extensively in the woodwind section. The value of these playing techniques as a
source of tension lies in their small-scale level of pitch activity in a harmonic context
that often contains sustained pitches and a slow rate of harmonic change. The flutter
of a trill and the timbral subtleties of a change in bow position add instability and
liveliness to pitch material; the slow, extended string glissandi in Du cristal
contribute to tension through producing motion and anticipation of their final
harmonic destination, and through the unique microtonal harmonic hues that result
from such incremental harmonic transitions.

The combination of parametric extremities that dominates climactic material across

the three works is rhythmic activity (unison) and dynamic profile. Registral spread
and harmonic density vary greatly with each appearance of rhythmic unison/high
amplitude climactic material, assumedly dependent on Saariahos desire to maximise
the climactic nature of the material. Wide registral spread is by no means strictly
attached to climactic material, with several examples demonstrating the striking
effect of condensing the registral spread of the ensemble to a small range see the
third section of Cendres (major 7th) or even the climactic unison A4 that concludes
the sixth section of Du cristal (the rhythmic unison in this case involves tied, long
note values across several bars).

The final section of Cendres is the only example of Saariaho using flux between
contrasting tensional states itself as a source of tension, releasing the work from the
more measured trajectory-based navigation of tension and release. As mentioned in

! 86!
the analysis of Cendres, while the work can be divided into sections based on tension
and release (allowing an overall formal structure to be perceived), it displays less
gradual tensional transitions than Nympha and Du cristal. Cendres displays a more
occurrences of several parametric changes occurring in parallel and demonstrates
Saariaho diversifying her methods of generating tension, with Cendres final section
being the most extreme example of this approach.

Parametric synchronisation is the main device through which points of arrival are
created in the three selected works. The impact of synchronisation in these works is
due, of course, to the vast amount of parametric desynchronisation that dominates
them individual players often have dynamic surges that are independent of the
rest of their ensemble, one string player will be playing sul pont while others play
naturale, several players will be performing rhythmic material that uses differing beat
subdivisions, and so on. Saariaho uses parametric synchronisation to bring the works
into sharper focus for high-tension passages by coordinating parallel changes in the
rhythmic activity and dynamic profile parameters specifically, with the addition of
the sound/noise parameter in the case of Nympha. All players will share a
crescendo, or align themselves with the same repeated semiquaver pattern, or, in the
case of Nympha, use extreme bow pressure for a type of timbral tutti. The works
then dissolve back into desynchronised webs of arrhythmic material with lone
dynamic surges that briefly push a player above the global texture before it is
reabsorbed. Synchronisation, of course, provides a final sense of resolution in Du
cristal using rhythmic unison, with the bass drum and crotales beating out of step
with one another until the final bars of the work, in which their final two beats are

Foreground ornamentation differs widely between the three works, both in terms of
how often it is used and the parameters that are used to provide it. While the
sound/noise parameter and the rhythmic activity parameter provide foreground
ornamentation for much of Nympha, there is less use of foreground ornamentation
in Du cristal, which relies purely on rhythmic activity, and very little in Cendres. The
key question regarding the use of foreground ornamentation in these works is why it
is used, or not, in a particular section or work. It seems that an ensembles size and
timbral homogeneity play a role in Saariahos use of foreground ornamentation, as
does her desire to emphasise periods of stasis or activity.

The statement that Cendres contains very little foreground ornamentation is

explainable due to the criteria mentioned above regarding ensemble size and timbral

! 87!
homogeneity. Cendres main difference from Nympha and Du cristal stems from its
relative restlessness: the active quality of its quick alternation between foregrounded
instruments, and lack of focus on creating globalised sound mass (although its
opening bars, and brief snippets thereafter, give a nod to the fused textures that
typify spectral music). The quick alternation of foregrounded material occurs
between instruments with distinct timbral differences piano, cello and flute
providing moment-to-moment variation of timbral foregrounding, occasionally
involving slight changes to the sound/noise parameter through the flutes breath
tones or a change of the cellos bow position. Cendres overall form simply makes less
use of stasis than Nympha and Du cristal. Its smaller ensemble size and quick
alternation between foregrounded instruments allows for less separation between
the micro- and macrostructural.

As touched on earlier, the rhythmic activity parameter is used in varying ways in

these works. Some passages in Du cristal contain polyrhythmic material as
foreground ornamentation, though Saariaho often places this material in the treble
register and above (such as the pitched percussion material in the latter half of the
third section), which keeps it from commanding any kind of temporal momentum,
but also moves the overall material away from stasis through providing active
foreground material. The use of surges of extreme bow pressure by individual
players in sections of Nympha serves a similar purpose, providing moment-to-
moment interest by breaking away from timbral homogeneity of the string quartet.
Points of arrival in Nympha and Du cristal are also of interest for the way that some
of this material blurs the role of foreground ornamentation, making it a source of
tension, with the parameter providing the foreground ornamentation being pushed
to its high-tension extremity. This is of interest because, while supplying moment-to-
moment interest, foreground ornamentation is not often the primary source of
tension. This blurring of roles is observed with the sound/noise and rhythmic
activity parameters in Nymphas climactic material. Elsewhere, the previously
mentioned percussion material of Du cristals climactic fifth section sees surface
material also play this dual role by becoming the primary source of tension.

To conclude, it is clear that Nympha, Du cristal and Cendres can be divided into
sections based on tensional processes that illuminate their overall formal structures.
The parametric changes that combine to create these tensional processes vary, both
between the works and within each work. Inharmonicity and the rate of harmonic
change were initially the two parameters given the most attention when analysing
the tensional processes in these works, primarily due to the primary points of

! 88!
difference in spectral and post-spectral composition being the use of slowly
developing timbre chords and non-tonal harmonic character. Analysis of these works
by Saariaho, however, reveal an equal or even greater reliance on rhythmic
changes for creating moments of high tension. This is most obvious through the use
of rhythmic unison in all three works to create points of arrival, providing temporary
perceptibility of pulse. Rhythmic unison always occurs in tandem with loud
dynamics that reinforce the significance of the material. Inharmonicity is also
common to high-tension material in each work, though not with the same regularity
as rhythmic unison and loud dynamics. As these works rarely shift out of
inharmonic chordal material (particularly Nympha and Du cristal), it seems that
Saariaho makes use of rhythmic unison as a further device for accentuating climactic
material. Parametric synchronisation is also a device that links the three works, as
several parameters will move to high-tension extremes simultaneously for climactic
passages, which also involves all players moving in parallel. The dominance of
desynchronisation of both parameters and players in these works again makes
for a contrasting point of arrival when synchronisation occurs.

A slow rate of harmonic change is common across all three works as a means of
supporting low-tension passages, while Cendres is the only work that uses an
accelerated rate of harmonic change to build tension. While at least a small element
of noise is present in each work, Nympha is the only work that truly harnesses noise
as a source of tension. Although fleeting, Cendres is the only work to embrace
harmonic consonance, allowing for more audible contrast in harmonic tension.
Regarding foreground ornamentation, Nymphas proto-melodic material and Du
cristals polyrhythmic material share the use of rhythm as foreground ornamentation
that subverts the underlying meter enough to avoid contributing a sense of

Stylistically, Saariahos organic approach to musical form, based not on literal

repetition but on tensional states and trajectories and variety in parametric
dominance, link her with other post-war composers that developed dense
polyphonic textures from which details could emerge, such as the vast polyphonic
and contrapuntal webs constructed by proto-spectral composers such as Ligeti. The
rarity of stable pitch and rhythmic material evokes organicism of the kind espoused
by Wishart (1996)22, who writes of the tendency for lattice-oriented music to be

! 22!Wishart, Trevor. On Sonic Art. Vol. 12. Amsterdam: Harwood Academic

Publishers, 1996.

! 89!
produced by composers in the West, with notated music fixed to the crossing lines of
equal-temperament pitch and discrete durational units.

! 90!

To conclude this exegesis, the final section acts as a segue between the subjects
covered in the analysis and the original compositions submitted as part of my
accompanying portfolio. I will point to aspects of Saariahos compositional
approaches that I have incorporated in the writing of the three works in my
portfolio, String Quartet No. 3, The Way Out Is Through for orchestra, and Ykai for
guitar. As stated in the Introduction, spectral music is a stylistically broad approach,
where the process by which compositional materials are derived is as much of a
defining feature as any kind of similarity between the works that result from this
process. I will give a brief description of the processes used to derive my
compositional materials, with reference to phenomena such as parametric
interaction, formal clarity, foreground ornamentation and the varying degrees of
spectralist compositional method in each work.

String Quartet No. 3

The first of the three works written for my portfolio, String Quartet No. 3 works with
the distortion of harmonic spectra. In this case the harmonic spectrum used as a
starting point is the harmonic spectrum based C2, the lowest open string of the cello.
A range of new harmonic spectra were then generated through stretching or
compressing the constituent frequencies of the original spectrum by various co-
efficient values, but retaining C2 as the fundamental pitch. The work then proceeds
to move through spectra that are increasingly stretched, until the first climax is
reached. The passage of lower tension that follows is made up of a progression of
increasingly compressed spectra. The passage from the second climax through to the
end of the work consists of a progression of spectra that are once again increasingly

In establishing the overall form of String Quartet No. 3, I worked with a

parametricised plan as outlined by Saariaho (1987:107)23. The end result in String
Quartet No. 3, however, is a high amount of parallel parametric change, in which
many parameters simultaneously follow the same tensional trajectory. It is
interesting to note Cohen and Wagners findings regarding the perception of

! 23 !Saariaho, Kaija. Timbre and Harmony: Interpolations of Timbral Structures.

Contemporary Music Review 2, no. 1 (1987): 93133. doi:10.1080/07494468708567055.

! 91!
increased musical tension in relation to parametric change, whereby the cognitive
load and, by extension, tension put onto the listener is actually increased when
perceiving parameters moving in divergent tensional directions24. Perhaps Saariaho
subscribed to this approach with the greater independence of her parametric
trajectories. As observed in the analyses of Nympha and Du cristal in particular,
passages of parametric divergence are employed, producing the muted climactic
material found in some sections.

In the general absence of noise-based tensional processes in String Quartet No. 3, I

made a conscious effort to use harmonicity/inharmonicity as a dominant source of
tension and release in the work. This was achieved by extracting pitch collections
from each distorted spectrum that contained intervallic relationships of varying
harmonic tension. Also related to the use of inharmonicity in String Quartet No. 3 is
my employment of pitch instability (another featured tensional device found in the
analysed works), not only through timbral manipulation but through techniques
such as glissandi, wide vibrato and extensive trilling that keep the moment-to-
moment harmonic material in a near constant state of motion in passages of higher

Though not a new technique in the context of my previous approaches to timbral

character in my instrumental writing, Saariahos timbrally detailed string writing
also informed my approach to String Quartet No. 3. Though the timbral character of
my string quartet is generally purer than that of Nympha, I attempted to instill an
internal activity and sense of movement to the pitch material that does not move in a
traditionally contrapuntal fashion, by incorporating shifts in bow position.

I also incorporated proto-melodic material as a form of foreground ornamentation in

String Quartet No. 3, in the first violin part between bb.6473. At this point in the
work, harmonic density has decreased from the claustrophobic double stopping that
initially dominates the work, providing an opportunity for introducing some
contrasting performative techniques to the material. I aimed to create active, yet non-
teleological foreground ornamentation that sat above the spectral transition
gradually unfolding. Saariaho, of course, employs proto-melodic material as

! 24!Cohen, Dalia, and Naphtali Wagner. Concurrence and Nonconcurrence between

Learned and Natural Schemata:The Case of J. S. Bachs Saraband in C Minor for

Cello Solo. Journal of New Music Research 29, no. 1 (March 2000): 23.

! 92!
foreground ornamentation in Nympha for example, in the violins fragmented
ascending runs in the first section of the work.

The Way Out Is Through

for orchestra

While the pitch material of String Quartet No. 3 is derived from a series of distortions
applied to harmonic spectra, the pitch material for The Way Out Is Through comes
from a more literal transcription of inharmonic spectra evolving through time (Fig
3.1). The sound source is a controlled swell played on a 22-inch ride cymbal, in which
amplitude moves from very low to very high, with the cymbal then left to vibrate
until any remaining sound was virtually inaudible. Two microphones were used to
record the swell, with one microphone positioned above the cymbal and the other
positioned directly underneath. The use of two vantage points for capturing the
sound of the swell is due to the difference in the range of frequencies that project
from the upper side of the cymbal in comparison to the underside (the underside
projects lower frequencies far more strongly than the upper side). A spectral peak
analysis was then performed at 13 different sample points during both microphone
recordings, comprehensively covering the arc of the swell, with the ten strongest
frequencies of each sample analysis rounded to the nearest quarter-tone (any
frequencies that lay outside of the range of orchestral instruments was omitted).

Ex. 3.1. Pitch material used for The Way Out Is Through, extracted from spectral
peak analysis of a ride cymbal swell.

The form of The Way Out Is Through is clearly an extended arc-like tensional
trajectory, based on the energy profile of the cymbal swell, with a brief percussion

! 93!
interlude momentarily interrupting the process at its tensional peak. I employ
several techniques found in the analysed Saariaho works to provide microstructural
interest while long-term processes unfold beneath.

As with String Quartet No. 3, parallel parametric changes dominate the first section of
The Way Out Is Through, following a strict process that applies brief moments of
swelling to several parameters, in a microstructural swell figure that interrupts the
low-tension, more timbrally detailed material that dominates. Brief, parallel increases
in dynamic profile, registral spread and harmonic density comprise these swells,
along with a purer timbral quality, as the strings revert to naturale bow positions.
Even the low-tension material contains small-scale swells in dynamic profile, though
these swells are not synchronized across the orchestra. This material and the more
synchronized multi-parameter swells gradually grow in intensity over several
minutes to reach a tensional peak at the mid-point of the work (echoing the peak of
the ride cymbal swell).

Changes in registral spread in The Way Out Is Through echo that of the cymbal swell,
with the widest spread occurring at the peak of the swell (tutti material at bb.5674)
and pitches in the bass register dominating the decay of the swell. For climactic
effect, the first half of the work employs a gradual expansion in registral spread that
begins with a small cluster of mid-register pitches, omitting pitches in the bass
register at the onset of the swell, introducing them once the controlled expansion of
registral spread eventually encompasses them.

Again, while not a new technique in the context of my previous compositional

output, I have added short-term timbral changes to slowly developing material,
adding microstructural life and activity as described in the Introduction. This takes
the form of constantly changing bow positions in the string parts of The Way Out Is
Through. These oscillating bow positions often coincide with dynamic surges,
creating a more marked timbral transformation, with the string player having to
apply more bow pressure to achieve a given dynamic when playing in the somewhat
counterproductive sul tasto position, as found in the first half of the work.

I employ parametric synchronization of the whole orchestra as a point of arrival in

The Way Out Is Through, with the orchestra briefly uniting for three thick, high
amplitude chords at bb.6973, with all players entering on the same beat, before
sliding back into the staggered, arhythmic surges that form the foundation of the
work. Parametric synchronization, and particularly rhythmic synchronization, is the

! 94!
primary parametric device used by Saariaho to herald a significant formal point of
arrival in her works, particularly in Nympha and Du cristal.

Pitch instability is incorporated as a tensional device, particularly in the high-tension

final bars of the first section (bb.5160), with trilling, wide vibrato, fluttertongue and
glissandi featuring strongly. As mentioned in the description of String Quartet No. 3,
this tensional device found in each of the selected Saariaho works helps to keep
moment-to-moment pitch and harmony in a near constant state of motion in
passages of higher tension.

Snatches of non-teleological foreground ornamentation of the kind observed in Du

cristal are also employed in the latter half of The Way Out Is Through, with
polyrhythmic material in the pitched percussion and harp parts sitting atop
sustained string lines sliding to pitches within new harmonic fields, with more
extended durations applied to the subtle dynamic surges from the woodwind and
brass sections.

for guitar

In a departure from the analysis-based sourcing of pitch material used for the two
preceding works, the handful of harmonic fields used in Ykai were created by
adding pitches around those used for the works motivic material. One of the
objectives of writing Ykai was to work with pitches of fixed register, whereby each
pitch class within a harmonic field rarely appears in more than one register. This
approach can result in modal instability as any given octave range within a harmonic
field seems to suggest a different scale or mode. Ykai also differs from the other two
works in my portfolio for its exclusive use of equal-tempered pitches.

While the tensional processes in Ykai are predominantly driven by changes in the
dynamic profile and inharmonicity parameters, Ykai is the only portfolio work in
which I employed an audibly accelerating rate of harmonic change as a tensional
device. This occurs during the works final climactic passage, from bb.77101, with
its shifting between fragments or elaborations of earlier motivic material and
harmonic fields. Here, the harmonic flux in the final section of Cendres provided an
influence to embrace more agile harmonic restlessness.

! 95!
To provide moments of contrast to the fixed register pitch organisation in Ykai,
ascending and descending runs found throughout the work incorporate pitches and
pitch transpositions from outside the underlying harmonic field. This device is
observed in Cendres, particularly in the pianos ascending runs that seem to act as a
type of chromatic refresher, or contributor of additional harmonic tension,
superimposed over the underlying harmonic field.

Form is perceivable in Ykai as much through literal repetition of motivic material as

through tensional trajectories, a departure from the dominance of tensional
trajectories in the perception of form in Nympha and Du cristal and, again, more
reminiscent of Cendres, which employs some literal repetition of material.

! !

! 96!

Berry, Wallace. Structural Functions in Music. Englewood Cliffs, N.J: Prentice-Hall, 1976.

Cohen, Dalia, and Naphtali Wagner. Concurrence and Nonconcurrence between

Learned and Natural Schemata:The Case of J. S. Bachs Saraband in C Minor for
Cello Solo. Journal of New Music Research 29, no. 1 (March 2000): 23.

Danner, Gregory. The Use of Acoustic Measures of Dissonance to Characterize Pitch-

Class Sets. Music Perception: An Interdisciplinary Journal 3, no. 1 (October 1, 1985):
10322. doi:10.2307/40285324.

Dibben, Nicola. The Perception of Structural Stability in Atonal Music: The Influence of
Salience, Stability, Horizontal Motion, Pitch Commonality, and Dissonance. Music
Perception: An Interdisciplinary Journal 16, no. 3 (April 1, 1999): 26594.

Farbood, Morwaread M. A Parametric, Temporal Model of Musical Tension. Music

Perception: An Interdisciplinary Journal 29, no. 4 (April 1, 2012): 387428.

Fineberg, Joshua. Spectral Music. Contemporary Music Review, 19, no. 2 (August 2000): 1
5. doi:10.1080/07494460000640221.

Granot, Roni Y., and Zohar Eitan. Musical Tension and the Interaction of Dynamic
Auditory Parameters. Music Perception: An Interdisciplinary Journal 28, no. 3
(February 1, 2011): 21946. doi:10.1525/mp.2011.28.3.219.

Howell, Tim, Jon Hargreaves, and Michael Rofe. Kaija Saariaho: Visions, Narratives,
Dialogues. Burlington, VT, Farnham, Surrey: Ashgate, 2011.

Huron, David. Interval-Class Content in Equally Tempered Pitch-Class Sets: Common

Scales Exhibit Optimum Tonal Consonance. Music Perception: An Interdisciplinary
Journal 11, no. 3 (April 1, 1994): 289305. doi:10.2307/40285624.

Johnson-Laird, Phil N., Olivia E. Kang, and Yuan Chang Leong. On Musical
Dissonance. Music Perception: An Interdisciplinary Journal 30, no. 1 (September 1,
2012): 1935. doi:10.1525/mp.2012.30.1.19.

! 97!
Lerdahl, Fred. Calculating Tonal Tension. Music Perception 13, no. 3 (1996): 31963.

Metzer, David. Musical Modernism at the Turn of the Twenty-First Century. Cambridge, New
York: Cambridge University Press, 2009.

Murail, Tristan. Spectra and Sprites. Contemporary Music Review 24, no. 23 (April 1,
2005): 13747. doi:10.1080/07494460500154806.

Plomp, R. & Levelt, W.J.M. Tonal Consonance and Critical Bandwidth. Journal of the
Acoustical Society of America 38 (1965): 548560.

Pousset, Damien, Joshua Fineberg, and Ronan Hyacinthe. The Works of Kaija Saariaho,
Philippe Hurel and Marc-Andr Dalbavie Stile Concertato, Stile Concitato, Stile
Rappresentativo. Contemporary Music Review 19, no. 3 (January 1, 2000): 67110.

Reigle, Robert, and Paul Whitehead (eds.). Spectral World Musics: Proceedings of the Istanbul
Spectral Music Conference. 1st ed. Istanbul: Pan Yaynclk, 2008.

Rose, Franois. Introduction to the Pitch Organization of French Spectral Music.

Perspectives of New Music 34, no. 2 (July 1, 1996): 639. doi:10.2307/833469.

Roy, Stphane. Functional and Implicative Analysis of Ombres Blanches. Journal of New
Music Research 27, no. 12 (June 1, 1998): 16584. doi:10.1080/09298219808570743.

Saariaho, Kaija. Timbre and Harmony: Interpolations of Timbral Structures.

Contemporary Music Review 2, no. 1 (1987): 93133. doi:10.1080/07494468708567055.

Sethares, W. A. Relating Tuning and Timbre. Experimental Musical Instruments 9, no. 2

(1993): 2330.

Wishart, Trevor. On Sonic Art. Vol. 12. Amsterdam: Harwood Academic Publishers, 1996.

! 98!

Saariaho, Kaija. Cendres. Wolpe Trio, Kaija Saariaho: Chamber Music. Kairos 0012412KAI
(2004) CD.

Saariaho, Kaija. Du Cristal. Los Angeles Philharmonic Orchestra/Esa Pekka-Salonen, Kaija

Saariaho: Works for Orchestra. Onedine ODE1113-2Q (1993) CD.

Saariaho, Kaija, Nympha. Kronos Quartet, CD. Kaija Saariaho Los Angeles Philharmonic
Orchestra, Esa-Pekka Salonen; Kronos Quartet; Petri Alanko, Anssi Karttunen Du cristal
la Fume / Nympha. Onedine ODE 804-2 (1989) CD.

! 99!

! 100!
String Quartet No.3

Brad Jenkins
String Quartet No. 3
Duration 830

Performance notes

Change bow direction ad lib during long tied passages. Players should not try to
synchronize bowing unless the ensemble shares a crescendo. All pitches connected
by a slur, including those that occur within double stops, should be played as legato
as possible.

All glissandi start at the beginning of the note on which they are notated.

All bowed tremolo are to be played as fast as possible.

Strong vibrato to be used, producing a wavering in pitch while performing


Touch harmonic pressure to be used on the string between beginning and end points
of glissando.
The Way Out Is Through
for orchestra

Brad Jenkins

The Way Out Is Through
Duration 9


3 flutes (1 piccolo double)

3 oboes
3 clarinets in B-flat
3 bassoons (1 contrabassoon double)

4 horns in F
3 trumpets in C
2 tenor trombones
bass trombone

3 percussion
(32, 28" and 25" timpani, marimba, glockenspiel, vibraphone,
2 suspended cymbals, triangle, bass drum, tam-tam)


Performance notes


Transposed score

Play the notated pitch a quarter-tone sharp.

Play the notated pitch a quarter-tone flat.

Trills on all instruments are to be of a semitone interval. It is at the players discretion as to

whether the trill uses the pitch a semitone above or below the main pitch. For trills occurring
on quarter-tones, players are free to use either a quarter-tone or semitone trill.

All glissandi start at the beginning of the note on which they are notated.

Where glissando lines do not lead to a destination pitch, the player is free to decide on the
destination pitch of the glissando, though the glissando should continue for the entire notated
duration and not stop prematurely. Trombonists may want to give themselves maximum
possible slide length for glissandi by playing starting pitches at the lowest available position.


Marimba is to be played with soft and hard mallets as indicated.

Tremolo in the marimba, timpani, tam-tam and bass drum parts indicate unmeasured rolls.

The two suspended cymbals should be of differing sizes, so that a timbral difference between
the two is audible.

Suspended cymbals are to be played with mallets.

Tam-tam is to be played with standard tam-tam beater.

Triangle is to be played with a standard beater.

Beater for the bass drum is at the players discretion.

Glockenspiel is to be played with brass mallets.

Mallets for the vibraphone are at the players discretion.

The indicated dynamic expression is to be achieved by quickly grabbing the cymbal with the
hand following the cymbal swell so that the cymbal continues to sound but the amplitude of
the swell is quickly reduced. An alternative to this method is to use the mallet heads to
dampen the cymbal.


Player is to gradually change between the indicated bowing positions.

All bowed tremolos are to be played as fast as possible.

Play with very wide vibrato for the duration covered by the vibrato line.


Moving bisbigliando glissando - play several fast glissandi one after another with alternating
hands. The fingers slide over the indicated interval as it moves upwards in register.

Thunder glissando - a single fast and loud glissando played by the left hand on the lowest
strings of the harp. The strings are played forcefully so that they rattle against one another
creating a metallic thunder-like sound.!!
! 138!
for guitar

Brad Jenkins
Duration 6

Performance notes

polp. polpastrella - angle the thumb to pluck or strum with the flesh alone,
producing a darker sound with considerable body.

rasg. rasgueado - strum with successive fingers.

Chords that are not preceded by an arpeggio line should be plucked with the fingers
together (chords marked rasg. are, of course, an exception).

The player should take care to perform portamenti and glissandi as distinctly
different types of slide. While glissandi begin immediately on their starting pitch,
portamenti should allow their starting pitch to sound for its entire (or almost entire)
notated duration before sliding to the next pitch.

Two identical pitches accompanied by two string numbers indicates that this pitch is
to be played on these two strings simultaneously.

The player is to perform a percussive slap with the fingers onto the strings and
momentarily hold the fingers against the strings, completely stopping any sound.
The pitches are to be played as fast as possible as an unmeasured, repeated arpeggio,
always in the sequence indicated. The tempo of the piece remains unchanged, with
the note value above the bracket indicating the duration of the technique.

There are two instances in which the unmeasured arpeggio occurs during a tempo
change: bb.106-108 and bb.112-113. In these instances the arpeggio should become
slower, following the molto ritard instruction.

Fret numbers above chords indicate that the same chord fingering is to be
maintained, simply moving up or down the fretboard to the indicated fret.
CD tracklisting

1. String Quartet No. 3

Reading by the New Zealand String Quartet (Helene Pohl and Douglas
Beilman, violin; Gillian Ansell, viola; Rolf Gjelsten, cello) at the Adam
Concert Room, Victoria University of Wellington, on 4 October 2013.

2. The Way Out Is Through

Reading by the NZSM Orchestra, conducted by Vince Hardaker,
at the Adam Concert Room, Victoria University of Wellington, on 22 May 2014.
NB: Instruments unavailable for recording: Flute 3, Oboe 3, Clarinet 3, Bassoon 2 & 3,
Horns in F 1-4, Trumpet 3. Contrabassoon added to recording via MIDI

3. Ykai
Performed by Jamie Garrick at the Adam Concert Room,
Victoria University of Wellington, on 27 September 2014.

! 148!